TL;DR: AI Inference hardware accelerators were all the rage a few years ago. They still are, but they seem to have abandoned the hobbyist, low-power, low-size, low-mid cost, seperate board user, such that abandoned projects such as the Google edge TPU from 2019 (5 yrs ago) are still your best bet $/perf wise. The $20 - $150 range is empty or has some products that aren't worth it at all. What happened? Are there any modern hobbyist $20 - $150 accelerators you can buy right now anywhere?
Sidenote: I know TOPS isn't the end-all be-all of perf comparison, but it's all I got.[1]
Skip for history of my interest: I've long been interested in machine learning, especially artificial neural networks since I took a class on ML in college in around 2004. I've done some hobbyist projects on the CPU and even released a C#/.Net wrapper for FANN (Fast Artificial Neural Network, a fast open-source neural network run on CPUs because everything was on CPUs then): https://github.com/joelself/FannCSharp. When deep learning took off I got excited. I got into competitive password cracking and although my ML based techniques were about a dozen orders of magnitudes slower at making guesses, they were almost immediately able to find a few passwords in old leaks that had been gone over and over for years by the best crackers with the most absurd hardware and extremely specially tuned password guess generators. That made me pretty proud that I was able to do something in a few months that years of dozens of groups with $100,000s of thousands of dollars of hardware and who know how many watt-hours couldn't do. I even thought about writing a a paper on it, but I was kinda in over my head and my life got a lot worse so I had to put all of my side projects on hold unfortunately. Recently though I did a vanity search for my FANN C# wrapper and found people talking about it and some references in some papers and student projects which made me feel proud.
Skip for history of my interest: Now I really want to get into the cross section of hardware-accelerated inference (no training this time, I'm not a trillion dollar company with billion dollars of supercomputers running on specialized training hardware that took 100's of millions of dollars to develop), microcontrollers for robots, drones, other smallish tasks that can't carry around their own 100 lb diesel generator and 2 1U rackmount servers full of inference hardware that I can't even get ahold of because you can only buy that stuff if you are an Intel or GE or some other company that might make products in the 10's of thousands at least. And this is where I hit a wall.
I just started looking around and one of the first things I found was Google's TPU by Coral.ai. 4 TOPs in a package, 2 chips on a small m2 card. Only about 40 bucks for developers to try out, $60 for an easier to use, but 1 chip only USB product. But this was about 5 years ago, and they just slowly disappeared and haven't made a peep in like 3 years. They timed the market perfectly. AI stuff was right on the verge of BLOWING THE F*CK UP. They could be THE edge/robotics/iot/anything-other-than-servecloud-phone-tablet-PC-laptop company. But they just seemed to give up. They're obviously not giving up on improving edge inference hardware. They release their phones twice a year (regular version, then A version) and they always update the tensor processing unit in those and are really starting to push that as a must have feature. They could use the same hardware improvements to make somewhat bigger chips to sell for other markets. You never know, someone might take their 3rd gen 16 TOPS TPU chip and makes a product(s) that takes the world by storm. Maybe multiple people/companies will do that.
Okay, so Google, seems to have dropped the hat. Hardware inference companies are a dime a dozen these days just go with another. But that's the problem. It seems all the focus is on Cloud scale, super-computer (some overlap between those 2), embedded on finished phones/tables/laptop/PCs, powerful server accelerators, and a very few extremely tiny MCUs with accordingly tiny MPUs. I seems everybody has abandoned the lower-mid range-robotics-drone-hobbyist space with haste.
ARM introduced the Ethos U-55 and U65 with the 65 having about double the TOPS of the 55 at a max of 1 TOPS in 2020. As far as I can tell the first products to use the U-55 were in 2022 and there haven't been a lot and I don't think they ran at top speed. Noone has opted to implement even an unmodified U-65 for anything. I recently bought a Grove AI Vision Kit with a U-55 NPU and it's specced at a lowly 50 GOPS (ARM's top-end says it could hit 10 times that and until *just now* I thought it was 500 GOPS and thus offered good $/TOPS ...oops).
There's a lot of companies making hype, a lot seeming to have selling dev or reference boards, but instead producing a few thousands and distributing them via the usual (Mouser, DigiKey, Element14, SparkFun, etc), they want you to fill out extensive forms to ensure you're a big player that will definitely eventually buy at least 100,000 units a day otherwise you're a waste of time for them to consider you (even though going over every applicant individually is WAY more time consuming than just producing a couple thou and have DigiKey take care of selling 1 to 2 at a time).
Thus I've come to the point that while Google edge TPU is abandoned (even though Google is going full steam ahead on AI inference for their cellphones and tablets) and Coral.ai is seemingly doing nothing. Their TPUs still provide the best $/TOPS in the range I want. Take a look at [VOXL2](https://www.modalai.com/collections/blue-uas-framework-components/products/voxl-2?variant=39914779836467). Basically exactly what I want and would expect we should have had something like a Google Edge TPU v3 by now (but a bit smaller and a little less power consumption, yes, I know moore's law doesn't really apply anymore, but in rapidly growing and learning fields like accelerated inference, double the speed every 2 years is not unreasonable and it has been 5 years since the Google TPU @ 4 TOPS per chip). But the damn thing is over $1,2000.
So, my point finally is that even though Google and Coral.ai seem to have abandoned their TPU. At about $40 for 2 chips at 4 TOPS apiece for 8 TOPS total, they still seem to be the best middle ground. The next best might be the BeagleBone reference studio for about 8 TOPS at $187. Same TOPS (though on one chip) for more than 4.5 times the cost. The Jetson Orin Nano by NVIDIA is $259 for 20 TOPS at $51/4TOPS that a single Google edge TPU will put out at $20 (including the board and stuff). It seems everyone is abandoning the hobbyist edge inference space at lightning speed. There are a lot of companies with promising size (physical) and performance products, but they won't talk to you until you fill out a form that implies that they only want to talk to someone that has already decided to buy 100,000s of your units whereas in the past, companies would have dev/reference boards out trying to find someone that would develop that killer app and make them a lot of money.
Why is this? Am I looking in the wrong place? Should I hoard Google edge TPUs? I bought their USB version to tinker with and the Grove AI Vision Kit (now that I realize is only 50 GOPS, so might be worthless). What are my options. For example. A single quadcopter a 100 - 300m above the ground looking "things", not image classic image classification where it can identify thousands of different objects. It just needs to identify one type of thing. Doesn't even have to be very fast. In fact, don't these NNs run on single images? I could just buy multiple chips and run in parallel to get the framerate I want if it isn't fast enough (it won't improve latency, but 100 - 500 ms latency probably isn't a problem until you get real close at which time you can switch to a different, much cheaper solution that works even better at close range and wide FOV).
Maybe I can use a phone and get low level access to the NPU/TPU and use that or use their powerful graphics cards on the phone or small laptop like a caveman from 2017. Still pretty expensive and I would be paying a ton of money for hardware I don't want. Maybe I could buy broken phones "for parts" on ebay, but I'm not that hardware savvy. I need a dev board to get me going.
The next best idea is to just push video from my drone/robot/project to a central station with a super powerful 1-4U server inference accelerator (not sure how I would get one), or Jetson Orin, computer with RTX4090 and do inference there and just tolerate the latency. That won't be feasible for some applications I would like to do though.

1. I found a github repo that collects perf comparison projects and I checked their data, and it's extremely sparse. One set is dominated by NVIDIA 4090, L(s), L4(s), and QUALCOMM T100 (or something, it's a cloud only processor, so you can't buy it). Then a few rows at the bottom have Raspi 4 and like 5 other mini applications units and MCU chips. And the results were hard to interpret especially since not all entrants have run all benchmarks and they can run it in probably dozens of different ways and then the results may not matter because their accuracy might have been bad. TOPS right now is like Whetstone/Drystone or MIPS, FLOPS, etc back in the day. It's a very rough estimate, but it can get you in the ballpark, so you can narrow down hundreds of options to 15 or so and then do more research from there. If someone comes up with something better then for sure let's all use that. Or if we could get some standardized benchmarks (I found some last night, there were several, and the results were very sparse (not every entry ran all of the benchmarks in all the different ways it could), one dataset was mostly a couple hundred rows NVIDIA 4090, L4, L40, QUALCOMM AI 100 (a cloud processor, you can't buy and run it) and then the last several rows where like a few Raspi 4s and some other MPU boards with drastically lower scores. Every once in a while some announces a project to fix this, but it hasn't helped at all.

TL;DR: AI Inference hardware accelerators were all the rage a few years ago. They still are, but they seem to have abandoned the hobbyist, low-power, low-size, low-mid cost, seperate board user, such that abandoned projects such as the Google edge TPU from 2019 (5 yrs ago) are still your best bet $/perf wise. The $20 - $150 range is empty or has some products that aren't worth it at all. What happened? Are there any modern hobbyist $20 - $150 accelerators you can buy right now anywhere?
Sidenote: I know TOPS isn't the end-all be-all of perf comparison, but it's all I got.[1]
Skip for history of my interest: I've long been interested in machine learning, especially artificial neural networks since I took a class on ML in college in around 2004. I've done some hobbyist projects on the CPU and even released a C#/.Net wrapper for FANN (Fast Artificial Neural Network, a fast open-source neural network run on CPUs because everything was on CPUs then): https://github.com/joelself/FannCSharp. When deep learning took off I got excited. I got into competitive password cracking and although my ML based techniques were about a dozen orders of magnitudes slower at making guesses, they were almost immediately able to find a few passwords in old leaks that had been gone over and over for years by the best crackers with the most absurd hardware and extremely specially tuned password guess generators. That made me pretty proud that I was able to do something in a few months that years of dozens of groups with $100,000s of thousands of dollars of hardware and who know how many watt-hours couldn't do. I even thought about writing a a paper on it, but I was kinda in over my head and my life got a lot worse so I had to put all of my side projects on hold unfortunately. Recently though I did a vanity search for my FANN C# wrapper and found people talking about it and some references in some papers and student projects which made me feel proud.
Skip for history of my interest: Now I really want to get into the cross section of hardware-accelerated inference (no training this time, I'm not a trillion dollar company with billion dollars of supercomputers running on specialized training hardware that took 100's of millions of dollars to develop), microcontrollers for robots, drones, other smallish tasks that can't carry around their own 100 lb diesel generator and 2 1U rackmount servers full of inference hardware that I can't even get ahold of because you can only buy that stuff if you are an Intel or GE or some other company that might make products in the 10's of thousands at least. And this is where I hit a wall.
I just started looking around and one of the first things I found was Google's TPU by Coral.ai. 4 TOPs in a package, 2 chips on a small m2 card. Only about 40 bucks for developers to try out, $60 for an easier to use, but 1 chip only USB product. But this was about 5 years ago, and they just slowly disappeared and haven't made a peep in like 3 years. They timed the market perfectly. AI stuff was right on the verge of BLOWING THE F*CK UP. They could be THE edge/robotics/iot/anything-other-than-servecloud-phone-tablet-PC-laptop company. But they just seemed to give up. They're obviously not giving up on improving edge inference hardware. They release their phones twice a year (regular version, then A version) and they always update the tensor processing unit in those and are really starting to push that as a must have feature. They could use the same hardware improvements to make somewhat bigger chips to sell for other markets. You never know, someone might take their 3rd gen 16 TOPS TPU chip and makes a product(s) that takes the world by storm. Maybe multiple people/companies will do that.
Okay, so Google, seems to have dropped the hat. Hardware inference companies are a dime a dozen these days just go with another. But that's the problem. It seems all the focus is on Cloud scale, super-computer (some overlap between those 2), embedded on finished phones/tables/laptop/PCs, powerful server accelerators, and a very few extremely tiny MCUs with accordingly tiny MPUs. I seems everybody has abandoned the lower-mid range-robotics-drone-hobbyist space with haste.
ARM introduced the Ethos U-55 and U65 with the 65 having about double the TOPS of the 55 at a max of 1 TOPS in 2020. As far as I can tell the first products to use the U-55 were in 2022 and there haven't been a lot and I don't think they ran at top speed. Noone has opted to implement even an unmodified U-65 for anything. I recently bought a Grove AI Vision Kit with a U-55 NPU and it's specced at a lowly 50 GOPS (ARM's top-end says it could hit 10 times that and until *just now* I thought it was 500 GOPS and thus offered good $/TOPS ...oops).
There's a lot of companies making hype, a lot seeming to have selling dev or reference boards, but instead producing a few thousands and distributing them via the usual (Mouser, DigiKey, Element14, SparkFun, etc), they want you to fill out extensive forms to ensure you're a big player that will definitely eventually buy at least 100,000 units a day otherwise you're a waste of time for them to consider you (even though going over every applicant individually is WAY more time consuming than just producing a couple thou and have DigiKey take care of selling 1 to 2 at a time).
Thus I've come to the point that while Google edge TPU is abandoned (even though Google is going full steam ahead on AI inference for their cellphones and tablets) and Coral.ai is seemingly doing nothing. Their TPUs still provide the best $/TOPS in the range I want. Take a look at [VOXL2](https://www.modalai.com/collections/blue-uas-framework-components/products/voxl-2?variant=39914779836467). Basically exactly what I want and would expect we should have had something like a Google Edge TPU v3 by now (but a bit smaller and a little less power consumption, yes, I know moore's law doesn't really apply anymore, but in rapidly growing and learning fields like accelerated inference, double the speed every 2 years is not unreasonable and it has been 5 years since the Google TPU @ 4 TOPS per chip). But the damn thing is over $1,2000.
So, my point finally is that even though Google and Coral.ai seem to have abandoned their TPU. At about $40 for 2 chips at 4 TOPS apiece for 8 TOPS total, they still seem to be the best middle ground. The next best might be the BeagleBone reference studio for about 8 TOPS at $187. Same TOPS (though on one chip) for more than 4.5 times the cost. The Jetson Orin Nano by NVIDIA is $259 for 20 TOPS at $51/4TOPS that a single Google edge TPU will put out at $20 (including the board and stuff). It seems everyone is abandoning the hobbyist edge inference space at lightning speed. There are a lot of companies with promising size (physical) and performance products, but they won't talk to you until you fill out a form that implies that they only want to talk to someone that has already decided to buy 100,000s of your units whereas in the past, companies would have dev/reference boards out trying to find someone that would develop that killer app and make them a lot of money.
Why is this? Am I looking in the wrong place? Should I hoard Google edge TPUs? I bought their USB version to tinker with and the Grove AI Vision Kit (now that I realize is only 50 GOPS, so might be worthless). What are my options. For example. A single quadcopter a 100 - 300m above the ground looking "things", not image classic image classification where it can identify thousands of different objects. It just needs to identify one type of thing. Doesn't even have to be very fast. In fact, don't these NNs run on single images? I could just buy multiple chips and run in parallel to get the framerate I want if it isn't fast enough (it won't improve latency, but 100 - 500 ms latency probably isn't a problem until you get real close at which time you can switch to a different, much cheaper solution that works even better at close range and wide FOV).
Maybe I can use a phone and get low level access to the NPU/TPU and use that or use their powerful graphics cards on the phone or small laptop like a caveman from 2017. Still pretty expensive and I would be paying a ton of money for hardware I don't want. Maybe I could buy broken phones "for parts" on ebay, but I'm not that hardware savvy. I need a dev board to get me going.
The next best idea is to just push video from my drone/robot/project to a central station with a super powerful 1-4U server inference accelerator (not sure how I would get one), or Jetson Orin, computer with RTX4090 and do inference there and just tolerate the latency. That won't be feasible for some applications I would like to do though.

1. I found a github repo that collects perf comparison projects and I checked their data, and it's extremely sparse. One set is dominated by NVIDIA 4090, L(s), L4(s), and QUALCOMM T100 (or something, it's a cloud only processor, so you can't buy it). Then a few rows at the bottom have Raspi 4 and like 5 other mini applications units and MCU chips. And the results were hard to interpret especially since not all entrants have run all benchmarks and they can run it in probably dozens of different ways and then the results may not matter because their accuracy might have been bad. TOPS right now is like Whetstone/Drystone or MIPS, FLOPS, etc back in the day. It's a very rough estimate, but it can get you in the ballpark, so you can narrow down hundreds of options to 15 or so and then do more research from there. If someone comes up with something better then for sure let's all use that. Or if we could get some standardized benchmarks (I found some last night, there were several, and the results were very sparse (not every entry ran all of the benchmarks in all the different ways it could), one dataset was mostly a couple hundred rows NVIDIA 4090, L4, L40, QUALCOMM AI 100 (a cloud processor, you can't buy and run it) and then the last several rows where like a few Raspi 4s and some other MPU boards with drastically lower scores. Every once in a while some announces a project to fix this, but it hasn't helped at all.

TL;DR: AI Inference hardware accelerators were all the rage a few years ago. They still are, but they seem to have abandoned the hobbyist, low-power, low-size, low-mid cost, seperate board user, such that abandoned projects such as the Google edge TPU from 2019 (5 yrs ago) are still your best bet $/perf wise. The $20 - $150 range is empty or has some products that aren't worth it at all. What happened? Are there any modern hobbyist $20 - $150 accelerators you can buy right now anywhere?
Sidenote: I know TOPS isn't the end-all be-all of perf comparison, but it's all I got.[1]
Skip for history of my interest: I've long been interested in machine learning, especially artificial neural networks since I took a class on ML in college in around 2004. I've done some hobbyist projects on the CPU and even released a C#/.Net wrapper for FANN (Fast Artificial Neural Network, a fast open-source neural network run on CPUs because everything was on CPUs then): https://github.com/joelself/FannCSharp. When deep learning took off I got excited. I got into competitive password cracking and although my ML based techniques were about a dozen orders of magnitudes slower at making guesses, they were almost immediately able to find a few passwords in old leaks that had been gone over and over for years by the best crackers with the most absurd hardware and extremely specially tuned password guess generators. That made me pretty proud that I was able to do something in a few months that years of dozens of groups with $100,000s of thousands of dollars of hardware and who know how many watt-hours couldn't do. I even thought about writing a a paper on it, but I was kinda in over my head and my life got a lot worse so I had to put all of my side projects on hold unfortunately. Recently though I did a vanity search for my FANN C# wrapper and found people talking about it and some references in some papers and student projects which made me feel proud.
Skip for history of my interest: Now I really want to get into the cross section of hardware-accelerated inference (no training this time, I'm not a trillion dollar company with billion dollars of supercomputers running on specialized training hardware that took 100's of millions of dollars to develop), microcontrollers for robots, drones, other smallish tasks that can't carry around their own 100 lb diesel generator and 2 1U rackmount servers full of inference hardware that I can't even get ahold of because you can only buy that stuff if you are an Intel or GE or some other company that might make products in the 10's of thousands at least. And this is where I hit a wall.
I just started looking around and one of the first things I found was Google's TPU by Coral.ai. 4 TOPs in a package, 2 chips on a small m2 card. Only about 40 bucks for developers to try out, $60 for an easier to use, but 1 chip only USB product. But this was about 5 years ago, and they just slowly disappeared and haven't made a peep in like 3 years. They timed the market perfectly. AI stuff was right on the verge of BLOWING THE F*CK UP. They could be THE edge/robotics/iot/anything-other-than-servecloud-phone-tablet-PC-laptop company. But they just seemed to give up. They're obviously not giving up on improving edge inference hardware. They release their phones twice a year (regular version, then A version) and they always update the tensor processing unit in those and are really starting to push that as a must have feature. They could use the same hardware improvements to make somewhat bigger chips to sell for other markets. You never know, someone might take their 3rd gen 16 TOPS TPU chip and makes a product(s) that takes the world by storm. Maybe multiple people/companies will do that.
Okay, so Google, seems to have dropped the hat. Hardware inference companies are a dime a dozen these days just go with another. But that's the problem. It seems all the focus is on Cloud scale, super-computer (some overlap between those 2), embedded on finished phones/tables/laptop/PCs, powerful server accelerators, and a very few extremely tiny MCUs with accordingly tiny MPUs. I seems everybody has abandoned the lower-mid range-robotics-drone-hobbyist space with haste.
ARM introduced the Ethos U-55 and U65 with the 65 having about double the TOPS of the 55 at a max of 1 TOPS in 2020. As far as I can tell the first products to use the U-55 were in 2022 and there haven't been a lot and I don't think they ran at top speed. Noone has opted to implement even an unmodified U-65 for anything. I recently bought a Grove AI Vision Kit with a U-55 NPU and it's specced at a lowly 50 GOPS (ARM's top-end says it could hit 10 times that and until *just now* I thought it was 500 GOPS and thus offered good $/TOPS ...oops).
There's a lot of companies making hype, a lot seeming to have selling dev or reference boards, but instead producing a few thousands and distributing them via the usual (Mouser, DigiKey, Element14, SparkFun, etc), they want you to fill out extensive forms to ensure you're a big player that will definitely eventually buy at least 100,000 units a day otherwise you're a waste of time for them to consider you (even though going over every applicant individually is WAY more time consuming than just producing a couple thou and have DigiKey take care of selling 1 to 2 at a time).
Thus I've come to the point that while Google edge TPU is abandoned (even though Google is going full steam ahead on AI inference for their cellphones and tablets) and Coral.ai is seemingly doing nothing. Their TPUs still provide the best $/TOPS in the range I want. Take a look at [VOXL2](https://www.modalai.com/collections/blue-uas-framework-components/products/voxl-2?variant=39914779836467). Basically exactly what I want and would expect we should have had something like a Google Edge TPU v3 by now (but a bit smaller and a little less power consumption, yes, I know moore's law doesn't really apply anymore, but in rapidly growing and learning fields like accelerated inference, double the speed every 2 years is not unreasonable and it has been 5 years since the Google TPU @ 4 TOPS per chip). But the damn thing is over $1,2000.
So, my point finally is that even though Google and Coral.ai seem to have abandoned their TPU. At about $40 for 2 chips at 4 TOPS apiece for 8 TOPS total, they still seem to be the best middle ground. The next best might be the BeagleBone reference studio for about 8 TOPS at $187. Same TOPS (though on one chip) for more than 4.5 times the cost. The Jetson Orin Nano by NVIDIA is $259 for 20 TOPS at $51/4TOPS that a single Google edge TPU will put out at $20 (including the board and stuff). It seems everyone is abandoning the hobbyist edge inference space at lightning speed. There are a lot of companies with promising size (physical) and performance products, but they won't talk to you until you fill out a form that implies that they only want to talk to someone that has already decided to buy 100,000s of your units whereas in the past, companies would have dev/reference boards out trying to find someone that would develop that killer app and make them a lot of money.
Why is this? Am I looking in the wrong place? Should I hoard Google edge TPUs? I bought their USB version to tinker with and the Grove AI Vision Kit (now that I realize is only 50 GOPS, so might be worthless). What are my options. For example. A single quadcopter a 100 - 300m above the ground looking "things", not image classic image classification where it can identify thousands of different objects. It just needs to identify one type of thing. Doesn't even have to be very fast. In fact, don't these NNs run on single images? I could just buy multiple chips and run in parallel to get the framerate I want if it isn't fast enough (it won't improve latency, but 100 - 500 ms latency probably isn't a problem until you get real close at which time you can switch to a different, much cheaper solution that works even better at close range and wide FOV).
Maybe I can use a phone and get low level access to the NPU/TPU and use that or use their powerful graphics cards on the phone or small laptop like a caveman from 2017. Still pretty expensive and I would be paying a ton of money for hardware I don't want. Maybe I could buy broken phones "for parts" on ebay, but I'm not that hardware savvy. I need a dev board to get me going.
The next best idea is to just push video from my drone/robot/project to a central station with a super powerful 1-4U server inference accelerator (not sure how I would get one), or Jetson Orin, computer with RTX4090 and do inference there and just tolerate the latency. That won't be feasible for some applications I would like to do though.

1. I found a github repo that collects perf comparison projects and I checked their data, and it's extremely sparse. One set is dominated by NVIDIA 4090, L(s), L4(s), and QUALCOMM T100 (or something, it's a cloud only processor, so you can't buy it). Then a few rows at the bottom have Raspi 4 and like 5 other mini applications units and MCU chips. And the results were hard to interpret especially since not all entrants have run all benchmarks and they can run it in probably dozens of different ways and then the results may not matter because their accuracy might have been bad. TOPS right now is like Whetstone/Drystone or MIPS, FLOPS, etc back in the day. It's a very rough estimate, but it can get you in the ballpark, so you can narrow down hundreds of options to 15 or so and then do more research from there. If someone comes up with something better then for sure let's all use that. Or if we could get some standardized benchmarks (I found some last night, there were several, and the results were very sparse (not every entry ran all of the benchmarks in all the different ways it could), one dataset was mostly a couple hundred rows NVIDIA 4090, L4, L40, QUALCOMM AI 100 (a cloud processor, you can't buy and run it) and then the last several rows where like a few Raspi 4s and some other MPU boards with drastically lower scores. Every once in a while some announces a project to fix this, but it hasn't helped at all.

I wanted to share this from a conversation about menopause information not being transmitted generationally:
For anyone who had this knowledge withheld, deliberately or not, you have a right to feel betrayed and angry. You are safe to express it here and should expect validation. Not arguing over anyone’s lived experiences.
We need to believe people from families who neglected them or worse, put the health of family members at risk.
I know each of our stories is different and complex, and always worth telling & listening to if we need support.
And please believe survivors of neglect & abuse who share their stories. Health issues can be traumatic & harder for people with previous traumas. 🙏
(If I listed the passive ways my mom endangered or threatened my health, and even my life, it would sound fake or like Grimm’s fairy tales. I get it.)

Ok, I don't think I'm an asshole but I tried to conform to the sub format.
I like to read human interest stories on Reddit, but my enjoyment is marred by the suspicion that most of the more popular posts are fake and even AI generated. It's always the same situations, very often written in a generic neutral style that lacks character. People rarely write this way on social media. They use slang, acronyms, colloquial expressions. They make mistake, or their mastery of English is limited, or they get incoherent because they're emotional.
About a month ago, I had a hunch and I logged in ChatGPT. I asked it if he knew what AITAH meant. He knew perfectly what it means and mentioned this very sub on Reddit. I asked it to write a fictitious AITAH post just for kick. He did. I then asked him if people ever asked it to write fake posts on Reddit.
THE FUCKING AI CRASHED! That question provoked a fatal error in the software!
Two weeks later, I learn in the news that OpenAI is investing hundreds of millions in Reddit to get access to their content and feed it to their machine learning robots.
I'm seriously worried that millions of Redditors are actually speaking to an algorithm. Am I an ass hole for thinking this? Actually, the question is more like Am I paranoid or what?
This post is not fake btw.

AI doesn't create anything new, it just mishmashes already existing things without much human input which makes it look "soulless". For now, because AI image generation is relatively new, the images look "good", but eventually they'll be the new GoAnimate or ClipArt. It doesn't help that AI image generator users tend to go for the same few styles.
Right now, today, and maybe for a few months, you're right! Most people would enjoy the media regardless of whether AI was used. Edgy people would say that all artists should be replaced with AI and those artists should get a "real job" at McDonald's.

Let's say that happened.

Every creator has been replaced with an AI equivalent.

You'll notice nothing at first, but eventually, you'll be tired of "anime" that all looks and moves the same. You'll be tired of music with the same chord progressions and lyrics with the same few themes. Every TV show, movie, book, and game will feel like the same experience (think Ninjabread Man and the other games made on the same engine).
You won't immediately think it's because of the AI, you'll just think the entertainment companies fell off, and you'll seek out indie and user generated content. What makes them more fun? Oh, right, the fact they're actually something new, and most importantly, made by people. The indie creators who use AI will seem as boring, or like they're copying the boring companies you lost all interest in.
If you like AI so much, why don't you just play games with AI instead of going online? Play TF2, Call of Duty, Overwatch, Smash Bros, etc. with bots only, since people are easily replaceable with AI. Oh what, are humans more fun? Interesting.
For a more extreme analogy, if you had a plastic bag on your head, you won't care since there's still oxygen inside and you can breathe, what a non issue! But eventually you'll be desperate for actually new air. Think of AI as a bag over the head of your interest in media, it'll die if you keep consuming what's literally the same exact content.

Tl;Dr You will be chronically bored of AI generated content and desperately seek human-made content. You WILL care if AI is used, just not because it's unethical.

/LonghornNation Daily Off Topic Free Talk Thread

Today: 5/29/2024

Last Thread

Current Austin Weather: 74° and Thunderstorm

Seven Day Forecast:

5/28	5/29	5/30	5/31	6/1	6/2	6/3
91°, Rain	89°, Rain	91°, Rain	94°, Rain	92°, Rain	89°, Rain	97°, Clouds

Your go-to place to talk about whatever you want. From the dumb shit aggies do on a near daily basis, to the latest whatever happening wherever. What ya got?

Here's a look at upcoming Longhorn Sporting Event(s):

1. 5/30 6:00 PM University of Texas Softball vs Stanford
2. 5/31 University of Texas Rowing vs NCAA Championships
3. 5/31 5:00 PM University of Texas Baseball vs Louisiana
4. 6/1 University of Texas Baseball vs Texas A&M/Grambling
5. 6/1 University of Texas Rowing vs NCAA Championships
6. 6/2 University of Texas Rowing vs NCAA Championships
7. 6/5 University of Texas Track & Field / Cross Country at NCAA Outdoor Championships

Trending on Reddit

/All

1. The amount of desk space that this pressure sensitive (“for your convenience”) mini bar takes up
2. Yann LeCun hitting Elon Musk right in his deepest insecurities
3. Are you certified
4. A car belonging to ISIS hits an anti-tank mine and gets yeeted several feet up before exploding
5. TIL that half of all plastics ever made were made in the last 15 years

/CFB

1. /CFB Donates $18,000.00 to Toys For Tots & Children's Hospitals, thanks to the 8th annual Holiday Drive!
2. Texas A&M AD Trev Alberts at SEC spring meetings: "We've just always had enough increasing revenue to overcome dumb expenses. I've said it 100 times, and I'll say it again: We don't have a revenue problem in college athletics, we have an expense problem."
3. [McMurphy] Texas' Steve Sarkisian on joining A&M in SEC: "There was so much talk about realignment. For us, we gained two rivals back. We're playing Arkansas & Texas A&M again. We're looking forward to it (A&M). I'm sure they are too."
4. [MattHayesCFB] Now at ⁦ @SatDownSouth : Sources: USC spent nearly 2 years trying to get out of playing LSU in Las Vegas. ⁩
5. Matt Baker - Auburn HC Hugh Freeze said Brian Battie is improving. But there’s a long road ahead for the former USF star. I’m sad to report that his prognosis, according to court records, is permanent paralysis.
6. [McMurphy] Texas A&M coach Mike Elko on Texas joining SEC: "We should play them. When you have 2 programs like that in same state 2 hours away they should play every year and it should mean a lot"

/LonghornNation

1. [5/28/2024] Tuesday's Sports Talk Thread
2. [5/28/2024] Tuesday's Free Talk Thread
3. 🥎 #1 Texas softball catcher and All-American Reese Atwood wins Player of the Year! 🤘
4. Texas Baseball will go to the Bryan-College Station Region as a 3 seed.
5. [POST-SERIES THREAD] 🥎 #1 Texas defeats #16 Texas A&M 2-1 and advances to the WCWS!
6. [5/27/2024] Monday's Sports Talk Thread
7. [5/27/2024] Monday's Free Talk Thread

LonghornBot: you can get a list of commands you can give for the bot by commenting ".help". You will receive a private message with the commands.
This thread was programmatically generated and posted on 5/29/2024 12:00 AM. If you have any questions or comments, please contact brihoang or chrislabeard

I’m not saying I’m the best DBD Mobile player out there (though I say I’m pretty good). But I’ve been around the block enough times to know that some survivors are just… special. You know the ones:
1.The “Locker Lizards”: These guys have a PhD in the art of hiding… and absolutely nothing else. They’ll sprint to the nearest locker the second the terror radius tickles their finely-tuned fear sensors and stay there, praying to the Entity for a swift and painless death (or at least until the killer finds someone more interesting to chase).
2.The “Follow the Leader… to Your Death”: These teammates think they’re leading a conga line, not escaping a bloodthirsty maniac. They’ll run you straight into the killer’s arms, then have the audacity to complain when you get hooked. Like, bro, I’m pretty sure I saw a “Do Not Enter” sign back there.
3.The “Skill Check… More Like Skill-LACKING”: These players could screw up a cup of instant ramen. They hit “good” skill checks about as often as I win the lottery (read: never). They’re basically a walking, talking generator explosion waiting to happen.
4.The “Every Man for Himself”: Altruism? Never heard of it. These survivors would rather watch you get sacrificed on a rusty hook than risk their precious hide to save you. They’re probably the same people who steal the last slice of pizza. And don’t even get me STARTED on the ones who bring a medkit and just…self-heal in the corner. Like, hello? I’m over here dying! Even worse? The ones who heal RIGHT IN FRONT OF YOU while you’re one hook away from seeing the upside-down world. Thanks, Claudette, I didn’t want to be rescued anyway.
5.The “Hold My Beer and Watch This” Daredevils”:These teammates have a death wish, plain and simple. They’ll attempt the most ridiculous, ill-timed jumps and pallet drops, only to end up face-down on the ground faster than you can say “Entity Displeased.” And guess who gets to play the hero? Ugh, fine, it’s me. It’s always me.
6.The “Escape is My Priority (Even if it Means You Die)”:Ah, yes. These lovely folks. They’ll bolt out the exit gate the SECOND it’s open, leaving you to the mercy of the killer. Extra points if they spam “Thank you!” as you’re getting chased by a chainsaw-wielding maniac. Real classy, guys.
7.The “But I Wanted to Totem!” Teammates: These players have their priorities SO messed up. Like, yes, cleansing totems is good… when the killer isn’t chasing us and we aren’t one generator away from escaping! Bonus points if they get downed right next to the totem and then spam “Help!” Dude, you HAD your chance!
8.The “Flashlight Fanatics (Who Can’t Aim)”:We all love a good flashlight blind, right? WRONG. Not when it’s done by someone who has the hand-eye coordination of a newborn sloth. They’ll blind you, the killer will turn around, and you’ll be left standing there like, “Thanks for nothing, buddy.”
9.“The Clicky McClickface”:This teammate has discovered the “tap to taunt” flash light button and apparently thinks it’s the only button worth pressing. Almost. All. Match. Long. Click, click, click. It’s like nails on a chalkboard, only amplified by the fact that they’re usually doing it RIGHT NEXT TO THE KILLER. Dude, I get it, you’re annoying. Now shut up and do a gen.

It doesn't make any sense. At all. Only Chujun and Kanako have boss monster DNA; and this must be the case because it is never hinted at in the game that Ceroba could be used for the experiment. It would have been 100% mentioned if she were.
And yet Ceroba, a supposedly completely normal monster with nothing special about her at all can:

1. Hit you with an undodgeable attack, a feat that has only been accomplished by boss monsters. Not even Sans can muster such a blatant FU to the game's rules.
   
2. Have insanely hard attack patterns, turning the game into a bullet hell. Even without the mask her attacks are way harder than those of normal monsters.
   
3. Generate a shield that MAKES HER IMPERVIOUS TO NORMAL ATTACKS. Clover has to turn upside-down to beat her.
   
4. Fly. In the animation going into the final phase you can clearly see her lift off the ground, and the sprite for that phase strongly suggests she is levitating, as you could not keep your balance standing that way.
   

How? This level of power flies in the face of all of Undertale and Undertale Yellow lore?

F27, My self esteem is taking hit after hit . I must be very ugly so I will not post pics to not ruin your day.
since i was 13 i was always on some sort of diet, some sort of cosmetic procedure forced by my mom. I put braces from 13 y.o till 17 y.o. Had skin peeling procedures at 15. Had a nose job at 19. Had a botox injection at 20. My hair is in bad state from weekly blow dry . 2 perms in the past 2 years. As well as a face liposuction consultancy. While being on a baby’s portion worth diet.with eating once to twice a day a very small portion.
Now she wants me to get a tooth shaving procedure for my bunny teeth and a retainer ( again) but this time permanent as well as veneers.
I must be very ugly, because when i say no she blames it on me for “ wanting to leave my faults unchecked “ She doesn’t fathom how i might not dislike my features as much. I even noticed she doesn’t feel hunger. She barely eats and tells me she always did. while everyone tells me i’ll fall apart if i don’t eat more. She is always bragging about how guys were heads over heels for her , and i am 27 and neverr dated so maybe she thinks if she fixes my appearance someone might like me.
As a child i wasn’t into makeup or cosmetics and only did it to please her. No i despise that fake plastic look, and i am not a super girly girly girl. I am more adventurous and kinda like exploring.
I know some replies might say that i am an adult but here your parents still have a say till you get married. It is different from the west. There are some cultural intonations to it as well.

With the surging success of Thrones of Decay and the recent positive news regarding Pharaoh, I'm honestly looking for a enjoyable project to ponder throughout the weekdays before heading home to work on my latest campaign. As such, I'm pleased to announce the introduction of The Newcomer's Guide to Army Composition, a multi-part series featuring effective generic and themed builds for each race within TWW3.
Every few days (ideally), I'll drop a different breakdown of a specific race's generic lord/hero/unit roster with effective, dynamic ways to fill out an army. Overall, these army composition recommendations hold to three core doctrines:

• Form follows function...these armies are designed to win battles, and the opportunity cost of utilizing a 'cool' unit over another 'less cool' unit means a great deal when we're talking about the defense of your empire. You know what's cool? Winning, and winning consistently and effectively.
  
• Cost-Effectiveness...we're not here to break the bank, and we're certainly not building 'doomstack' clones (seriously not fun). Great army compositions have layers with a variety of backfill options available...you can't always afford the highest level in every category for every army, so the name of the game is cost-effectiveness.
  
• Flexibility...if you take one thing away from this guide, I hope this tops out your list: variation, flexibility and durability are the saving graces of the best army compositions. I’ll admit that opponent specialization has certain advantages, but generally speaking, I despise the idea of sitting around for ‘X’ amount of turns globally recruiting the ideal niche unit designed to defeat [insert race here]. Instead, I find a force capable of challenging virtually every race much more useful in the mid-game and down the stretch of long campaigns.
  

Alright, on to the good stuff…first up, a personal favorite: the Dawi.
While I personally recruit based on the legendary lord and which advantages they provide certain units, whether that's through their unique trait or via their unique skill line, my goal with this breakdown remains typical army compositions for non-legendary lords...here's how I approach a relatively generic build:

• First and foremost, pick yourself a lord type...the generic lord is really solid, the runelord serves as more of a support specialist and the new daemonslayer offers a extremely different melee approach. Unless you're building a slayer-themed army, I'd generally suggest one of the first two as the daemonslayer truly works best when utilizing a good number of slayer units. Total Units: 1/20
  
• Next, and I've found this is an unpopular decision on this forum, I utilize a strict one hero per army rule with a few minor exceptions (VC, for example). I also offset my lord choice with a hero of opposite specialization...so a runesmith would pair well with a dwarf lord while a thane works perfectly with the runelord. The master engineer slots in well with both, so just modulate your needs accordingly. And if you're the type who prefers more heroes, throw in one of each, all three are excellent. I'll note that the new dragonslayer is also quite brilliant, really knows how to take on large entities and a great option to throw into the mix as well, especially if you're going to be fighting tomb kings, dark elves, greenskins, etc. Total Units: 2/20
  
• Six frontline infantry designed to hold the line...these traditionally are well armored and shielded and just generally stay alive. They aren't your damage dealers and they aren't designed to generate kills...they hold the line to ensure your actual killers have time and space to do their job. Start off with Dwarf Warriors (easily the best Tier 0 unit out there) before upgrading to Longbeards and, finally, Ironbreakers. The latter offer some excellent punch with their short-range explosive toss, excellent in the late game against heavily-armored troops such as Warriors of Chaos or the Dawi-Zharr. Total Units: 8/20
  
• Behind your frontline, you need two kinds of damage dealers: ranged and melee. First up, let's talk about the ranged units...Thunderers are the way to go, bar none. I'd argue both variants -- the traditional and the grudge-rakers -- are exceptional, with the latter maybe being slightly better due to a quicker reload time. The traditional have slightly better range and accuracy, though, so there's some trade-off. My go-to from now on will be two of each, to be honest. If you’re a bit strapped for cash, quarrelers are always a solid choice, plus you can run some simpler formations as those crossbows shoot up and over your frontline units. Total Units: 12/20
  
• Now you need some hard-hitting melee, and fortunately you have some options. Hammerers were the original choice on this front, but with the addition of doomseekers, it's a bit more of a toss-up. The question you have to ask yourself is survivability...hammerers, with their high armor and high numbers (100 per unit), will hold up in most fights. Doomseekers with their zero armor and lower numbers (32 per unit) don't have that advantage. That said, doomseekers have some sick animations and way outclass the hammerers in their general stats (better MA, MD and charge bonus), so it's still a bit of a toss-up. I bring two total to help my frontline with particularly tough enemies, and with both offering magical damage, ensure targets consistent of the undead or daemonic, when applicable. When you’re cost-strapped, look to the great weapon variants of long wards and dwarf warriors. Total Units: 14/20
  
• Not much of a Dawi force without some artillery...the usual approach on here tends to be an overwhelming amount in this category, but I honestly think two total truly works perfectly within most armies. Start off with grudge throwers before upgrading to cannons and then organ guns and flame cannons. Full honesty, I’m completely sold on the new flame cannons…they absolutely melt most units, and with a Rune of Seeking, full on chef’s kiss my friends. Pair with an accurate damage-dealer such as organ guns and you’ll have enemies severely crippled well before reaching your infantry frontline...that'll do nicely. Total Units: 16/20
  
• Moving on to war machines...these are the units you'll need to micro during battles to get the most impact, so just prepare for that inevitability. Prior to ToD and the 5.0 patch, the unit of choice was the gyrobomber, but I've honestly been forced to reassess a bunch the past couple weeks. First off, the Thunderbarge is a no-brainer, grab one of those as soon as you can (assuming you can afford it)...just float that sucker over the enemy army within range of your thunderers (for protection) and it'll take care of virtually anything. Second, the gyrobomber is still great, and it's a solid option for harassment, but I really appreciate the utility of the gyrocopter (brimstone guns)...they mop the floor with a heavy majority of opposing ranged options, freeing up your thunderers and Thunderbarge to focus on the opposing frontline. Mix-and-match between the gyrocopters and the gyrobombers for best results, but I've found just one with a thunderbarge to work really well. Total Units: 18/20
  
• Finally, the flex positions...I tend to slot in some anti-cavalry units for these two final spots, just to ensure my artillery doesn't get ravaged by boar riders or the like. Surprisingly, Bugman's Rangers work exceptionally well in this hole...they have this under-the-radar 'Charge Defense Against Large' ability that makes them ideal for taking on cavalry, and up until the point they get charged, they're a solid missle unit as well (same effectiveness as a quarreler). However, my favorite unit for these two spots are Irondrakes (Trollhammer Torpedoes)...they crush it against large entities, can hold their own in melee for a decent amount of time due to their armor and just generally look super cool. Total Units: 20/20
  

A Slayer-Themed Composition
As promised, I also included a great slayer build that I have come to enjoy tremendously:

• Daemonslayer for the lord slot…a must-have for this composition, feels strange to run anything else. Total Units: 1/20
  
• Dragonslayer in the hero position…again, a must-have, and truly excellent as a single-entity killer. Total Units: 2/20
  
• Doomseekers…prior to the addition of these beasts of the battlefield, slayers were entirely focused as anti-large specialists. Doomseekers changed that entire dynamic, and I recommend running six of these gems as your ‘frontline’ for this comp…a bit expensive, sure, but if you’re building themed armies, you’re already rich. Total Units: 8/20
  
• Giantslayers…still the anti-large champ in regards to slayers, and highly-effective on your flanks protecting your doomseekers from cavalry charges and monstrous infantry. I personally grab four, two on each end to ensure the integrity of my frontline. Total Units: 12/20
  
• Slayer Pirates…an extremely fun unit with some late-game purpose in this composition: ranged firepower with a bit of flanking capability. Like I noted earlier, I appreciate flexibility on the campaign map, and the last thing you want to run into with a themed stack is a couple enemy units virtually untouchable by your forces. Slayer Pirates solve that problem well as your anti-flying shield, and when that need isn’t as ‘needed,’ per se, they make solid flankers with excellent armor-piercing ranged ammo. I bring four for safety. Total Units: 16/20
  
• Goblin Hewers…not exactly my favorite of the slayer units, but definitely fun in the right context and against any lord boasting a large mount. I toss in a pair to absolutely delete enemy leadership or large single-entity monsters, and they’re the type of ‘artillery’ you never have to worry about since they’re manned by another couple dozen slayers…cavalry charges need not apply. Total Units: 18/20
  
• And finally, a few Gyrocopters (Brimstone Guns)…I know, I know, not technically slayers, and while you may be absolutely correct, the advantage of a pair of brimstone copters takes root in their capability taking out ranged units. Slayers, as we all know, sprint into battle sporting zero armor, making them vulnerable to even the sloppiest of archers. In steps a pair of brimstone guns, and suddenly you’ve got very little ranged to worry about plus some post-battle barbecue options for your surviving slayer comrades. Total Units: 20/20
  

Enjoy the Dawi, they are an absolute blast to play with now, and I highly recommend experimenting with some of their mid-tier units as well as virtually all variants of grudge settlers…the flame cannons and trollhammer copters are unreal out in the field.

Race Information

• Name: Vermont City Marathon
• Date: May 26, 2024
• Distance: 26.2 miles
• Location: Burlington, VT
• Website: https://www.runvermont.org/
• Age: 35M
• Time: 3:26:22

Goals

Goal	Description	Completed?
A	Sub-3:20	No
B	Sub-3:30	Yes
C	Finish	Yes

Splits

13.1 splits	Time
1	1:43:20
2	1:43:02

Ancient History

Growing up, I was always pretty active and started running in High School. Going into my Junior year, I began to take running much more seriously and made steady progress with a 5K PR of 17:06 at the State meet. Going into my Senior year, I had been training 6 days a week with the team and hitting 40-60 miles a week consistently with hopes of going low 16's by the end of the XC season. Then, disaster. Pain in my quads was getting worse and worse until a bone scan finally revealed bilateral femoral stress fractures - one midway along the midline of each femur. I was devastated, and completely stopped running from August until February. By then, it seemed like things were getting better and I was able to put in some good miles for the spring Track season. I was always a better mid-distance runner, and was able to snag a few PRs of 2:00 in the 800m (agony, 1s off) and 4:35 in the 1600m. By that point I was basically over running, completely burned out and sick of running.
10 years later...
I wake up one morning in April of 2017 to take a shower and realize that I have no temperature sensation in my right leg whatsoever. I'm now finishing grad school, I've moved across the country, first kid on the way and due in a matter of weeks. I'm not working out much at this point, maybe running every few months at best. Stress is high. I go to urgent care and the last 6 months of subtle pins and needles on the left side of my face and the occasional double vision start to make a lot more sense. Multiple Sclerosis. It's not a death sentence, but I felt like the world was ending. My kid was due in 4 weeks and now my future was completely uncertain. A month or two, an MRI and one spinal tap later and the diagnosis is confirmed.

Training

Fortunately, new medicines have made MS much more manageable and I was lucky to have a great medical team. Perhaps the most important thing to keep the disease in check beyond highly potent immunosuppressants is regular exercise. Time to start running again.
Starting out was rough, but within a few weeks I was able to feel ok running 15-20 miles a week in the 8-10 min/mile range. I keep that up consistently for a year or so and run my first 10K with a time of 44:42, which I was pretty proud of at the time. Things are going well for a while, second kid is on the way, my MS seems mostly stable, albeit leaving me with some permanent loss of function of my left leg (these gait issues will come up later), pins and needles in my right leg, and fatigue. Then COVID hits. Now I'm at home, with much more time. I increase my mileage a bit, now running 20-30 miles a week but not really following any particular plan or anything. Move across country, keep running when I can.
Mid-2023 I decided to finally take it up a level. I was mostly using the Garmin Daily Workouts at this point and running consistently 30 miles a week. I decided to run my first half-marathon, just as a virtual training run, and felt pretty good about my time with an 8:24/mi average pace. Going into fall, I juggled some illnesses but kept running. Unfortunately around November I started to develop Sesamoiditis and had to really limit my running to 35-40 miles per week. I ran a Christmas 5K with a time of 19:14, which felt pretty good, but was still dealing with the Sesamoid issue.
Over time, the Sesamoid started to resolve but I would occasionally feel some twinges in my right Adductor. Nevertheless, I felt like the time was now to try for my first Marathon. Around February, I finally committed and decided to do the Pfitz 12/55 plan. I had been consistently in the 35-40 range for a few months, so felt like I had a decent base. At first, I found the plan quite difficult to deal with. It was the most structured running I had done since high school, but after a while my schedule adapted and I was hitting all of the workouts. About halfway through, the Adductor issue started to really rear its ugly head. It seems like it's a combination of gait issues from my MS as well as compensating for the weakened Sesamoid. I attempted to strength train, but had to back off a bit.
At best, I was able to hit 53 miles a week, occasionally having to skip some workouts for cross-training or rest. I felt like I nailed the 20 mile runs pretty well and overall the LRs felt good. I ran one 5K as a kind of "tune-up" with a time of 19:21, pacing a friend, so putting in maybe 80-90% effort. Due to injuries and some travel I was never able to do one of the longer tune-up races. Unfortunately, the Adductor issue continued to wax and wane, finally flaring up badly about 3 weeks out from my target marathon. After a consultation with Ortho, I was told to stop running cold turkey for two weeks prior to the Marathon, and then a follow-up last Friday I was given the green light to try it ... cautiously ... but consider dropping out if it is too painful.

Pre-Race

My taper was compromised pretty heavily by the injury, so I was really feeling worried about how the race would go. That said, I knew that the training was done at this point, and so an extended rest shouldn't make TOO much of an impact if I feel ready. I carb-loaded three days out and tried to take on a lot of fluids. While my injury was feeling better, the next worry was the temperature. Forecast was saying low 60s for the start of the race, ramping up to the mid 70s by the time I expected to finish. We drove up to Burlington from the Boston area on Friday with the kids, and I likely did too much walking on Saturday but overall I was feeling ok and was fortunate to get a good amount of sleep Saturday night (7.5hrs).
I'm a scientist by training, so planning and preparation is second nature to me. I woke up at 4:45, had a cup of coffee, half a bagel, a banana, and a Maurten 160 drink. Because of the temperature, I decided to prepare two 500mL bottles of Tailwind, one that I would take with me and one that I would pickup from my wife at the Half point. I planned for 4 gels (Maurten@5mi, Gu@10mi, Maurten+Caf@15mi, Maurten@20) and to get a cup of water at every aid station each 2.5mi. I jogged down to the start at 6:15am, used the restroom probably 4 times, and lined up with the 3:30 group.

Race

My plan was to start out with the 3:30 pacer group and see how things go. The course is advertised as being flat and fast, but I'm not sure I believe that because there is a big hill you run up twice with 120ft of vert over about 1/2 mile and between 5-7% grade at times. The course is essentially two figure 8s, and you start in the middle. At 7:15am, we took off. The first few miles felt a bit stiff, but I was chatting it up with the pacers and feeling pretty relaxed. We were going a bit ahead of pace, clocking in at 7:50 per mile for the first 4-5 miles. Around the 10K mark, I was feeling a bit antsy and the pace was slowing down...I decided to head off alone, despite a lot of reservation that I was making a poor decision.
Around Mile 9, I was starting to feel a bit of tightness in my legs and my HR was in the high 160s. I was a little bit concerned about this early fatigue, but I knew from my training that I feel like this on almost every single run and it doesn't really seem to get worse, it just seems to be part of my mechanics or something. I kept pushing on, mile after mile, keeping my pace pretty consistently. Every aid station I got a cup of water, drank some and splashed the rest on my head. This made a HUGE difference in the end.
Mile 13, came in at the Half feeling great. Started up one of the hills and saw my wife. *Disaster* she had the wrong bottle. No carbs, no Tailwind for the 2nd half, and my current bottle was empty. At this point, I had also been taking my gels early. My stomach was feeling great so instead of a gel at every 5th mile I was taking one at every 4. The race provided gels at mile 8 and mile 21, so I had picked up an extra gel by this time and made the decision to stop at the next aid station around mile 15 and fill up my bottle with Gatorade. Salvation.
Kept chugging along, feeling a bit of fatigue setting in around the slog from mile 16-19. At mile 19, I saw my wife again and she had the correct bottle (it was my fault, she thought I meant for her to give me a larger, recovery bottle I had prepared of Skratch for after the race). I got my bottle of Tailwind and hit the monster hill at Mile 20. This was my slowest mile at 8:15 (GAP of 7:40) and it really sapped my strength. I was thinking that this must be like running Heartbreak Hill the entire time I was going up.
Through the hill, into the last 10K. Starting to feel desperate, just pushing forward every step of the way. The pacing in this marathon is quite strange, because it also has a Relay of either 2-person or 4-person teams, so you're always seeing different people with way fresher legs than you. Ended up finding a woman to run with who was crushing it, and we paced eachother the last 3-4 miles. Took a last gel around 24 miles and grinded it out to the end.

Berlin next... then Boston?

Post-race, I was feeling pretty happy with my 3:26 time. To be honest, I felt a little anti-climactic, although I'm not entirely sure why. I didn't feel a ton of adrenaline or emotion throughout the race and things seemed pretty collected. That said, I'm pretty surprised at how much I loved almost every minute of the race itself. Maybe it was just because it was my first time, but it was truly a fun and rewarding experience.
I think there were a few areas where I could have pushed just slightly harder, but given it was my first marathon on a somewhat challenging course in the heat I'm overall happy with my time. I absolutely nailed my hydration/fueling and my pacing, with a slightly negative split overall, so I'm really pleased with that. As a whole, I'd give my training cycle maybe a 6/10; I think I definitely raced too many of my training runs, likely leading to some of my injuries. I was also only able to go above 50mpw in 2 of the 12 weeks of the program.
Given my somewhat poor training cycle, I'm hoping that I still have a lot of room to improve. I was a bit worried that my MS would cause issues during the marathon, particularly my left leg, which tends to lose coordination in long, higher intensity efforts. Fortunately, it felt pretty good throughout.
Earlier in the year, I was lucky to get a spot for both Berlin and NYC. Given their proximity to each other, I'm going to try and defer NYC to next year. I've now been thinking through if I want to try and apply to Boston for 2025. Given my MS, I am technically eligible for an "adaptive" application, which has a cutoff time of 6:00. I feel in two minds about this, because on one hand I feel like I'm maybe taking a bit of the easy way out, when it might be possible for me to hit sub-3:05 some day. On the other hand, I don't know how many years I have left running so I'm thinking I might just seize the opportunity now and then try and qualify through the more "standard" way in the future. Curious on people's thoughts on this, and thanks for reading!
Made with a new race report generator created by herumph.

Triprolidine is a first-generation antihistamine, which was the antihistamine ingredient in the "old" Actifed (before they reformulated it). It is extremely sedating, in my experience (which I know first hand because I used to take Actifed for allergies). At least sometimes. Of all antihistamines, including hydroxyzine (which will make me very sleepy - the next day... Not when I need it to...), diphenhydramine, and doxylamine, this one can sometimes hit me like a ton of bricks when the others are largely not noticeable. When it does make me sleepy, it takes me from chipper and wide awake to "OMG, I absolutely can not keep my eyes open, I can not stay awake no matter how hard I try" within 30 minutes, plus or minus.
Does anyone know where this can be obtained, and if it can, what product name(s) it is? It's still approved for OTC use in the US, and I believe Walgreens still has their "Cold and Sinus D", which is supposed to be basically the same as the old Actifed formulation. But it (like Actifed before it) is in combination with pseudoephedrine, and I would assume pseudoephedrine is about the last thing you'd want to take if you were having problems getting to sleep. Is it available as a single ingredient product somewhere?
Even in the combination, whether it helps or not is spotty - sometimes it does, sometimes it doesn't. Can anyone speak to how to maximize the chance of getting the sleep inducing effect? Take it on an empty stomach? Take it with food? Something else? I can't say that I ever noticed a pattern.
I know that one should not lean on first-generation antihistamines as sleep aids, at least not for any length of time anyway (plus I'm sure it will lose effectiveness if taken regularly over a period of time). But man, it's the one thing I ever found that actually totally knocks me out (sometimes), and back then, that's not what I was intending to happen.... And now that I want to use it for that purpose, I can't find it...

I can't believe I'm writing this. But I need some insight cause I find myself getting angry and confused. This is my first reddit post. Please excuse my redditor literacy.
This is the most convoluted story. It is long. But it's a ride.
You've been warned.
(Fake names obviously)
I (20f) have a best friend, Karl (20m) of 4 years. Now I see what you might assume. No. We're close but I'm mainly into women, I currently have a partner and have had a partner 90% of the time they where dating.
Now Karl got with Regina(19f) late 2022, the relationship started off rocky as she said "I only want you" but then kissed her ex, and then couldn't decide who she wanted. But still insisted once she chose Karl, she wanted to stay friends with her ex. Posted pictures when they'd "hang out" where it looked like she was sitting on his lap. But she swore she wasn't. Constantly blocked him after things would happen, then unblocked him, lied, then cried when Karl would find out.
Yes. Infuriating. But here you go. That's how Regina was introduced to our lives.
It took a while but eventually I tried to look past this. I care about Karl, if this was who he loved. We accepted it. Infact made it a point to invite her out to gatherings, made sure to offer her food, offer her drinks, chatting. Making sure she's involved. Gassing her up. Girlie things. (God I'm so desperate for everyone to love me it's a problem.)
Then her friends, ex boyfriend began to follow me, I had hoped this was because of how well I'd done to make friends. But this waa short lived.
Originally I'd just hoped it was banter. I'd chat to them, often sending pictures with Regina in her classes and joking with me.
Unfortunately I have social impairments, Slowly it became clear they where just laughing at me, calling me names but with cutesy emojis. Remember the girls in highschool? The ones thatd pretend to be your friend in class because it was funny? Like that.
So i stopped paying attention, often ignoring them. Unfortunately it only got worse. It got to a point I'd be spammed and have my instagram story replies with slurrs, calling me a pdf. File??? (I was talking to someone 6 years older than me?) Weird references, calling me cringe (I know. I know, worst thing ever right.), picking on my hair, my eyes(strabismus), my clothing. So I folded. Told Karl I wouldn't be dealing with it anymore. I'd blocked them, and asked karl To ask Regina to ask her friends to stop contacting me, I was doing my finishing project in college (uk) to get into university and it was getting to point I couldn't focus. I told him what had been happening, that I didn't know what her problem was. But I am a adult woman and this was bizarre.
Now, that alone. I forgave and in time, forgot. She had allegedly appologized "for them" and didn't know any of that was happening and had no I'll intent and hoped we could still be friends. Okay, sure.
Weird semi important point: she confessed in a groupchat that she used to be a 'chav' I said " you do look like someone who'd have bullied me" Banter. She then posted on her Instagram story (Paraphrased by memory) "When someone says you look like someone who'd have bullied them- but your friend died" I can't remember, but it was along the lines of that kind of 'what the fuck does that even mean'
Upon a later night of drinking, regina was talking to Karl about the ex, Mike. I brang up the fact her ex boyfriend kept liking my photos and was following me Hoping to bond over the fact this guy was weird, common girly bonding
"You know he only follows you so he could make fun of you and how cringe your posts are". She laughs.
The group goes quiet and holy shit I'm embarrassed. I just internalise that and change the subject.
Later I repost a reel of a guy saying something vaugely corresponding to this convosation. Basic premise when someone tells you their friend talks shit about you, then obviously you ask "why do they do that to you" (I know childish but at this point I was starting to really dislike her. My friend had sent it to me, It was late.) When i say She launched, "if you've got a problem talk to me instead of being weird and I'd tell you I was so scared of Mike and he held such a power over me and I just let him chat shit" I'd love to just mention this is after the 2nd time she'd unblocked him to talk to him behind Karl's back. I put up with it. Karl is at this point family. And if this is who he loves. We have to love her too.
This is all important to the point I swear.
Anyway.
My partners (now ex) friend Frank (22) and us fell out. Unimportant to this story but he let me know, Regina and an old very close friend had a groupchat to say very unpleasant things about me in, despite this old friend I never stopped speaking well of. Hoping we'd find eachother again. He'd been scouted when we had fallen out. But respected me enough to tell me. Another confrontation where she is so misunderstood and I'm making a big deal out of nothing and she's never ever had a problem with me.
Okay. Talked to Karl again. He is shocked but takes her word. As I'd kinda expect. Its his girlfriend. He took her to London over my birthday, he didn't want to ruin it. So he gave it up.
Karl throughout this is withdrawing from us. When he's with us it's like the light is gone from his eyes. He's distracted, quiet, doesn't laugh as much. Often tries to slip out of meetups because he'll "only bring you guys down". He's constantly picking up his phone. Constantly messaging. Cancelling plans. He won't talk to us. We where all worried.
Karl few months later calls me for advice. Turns out she kept getting caught in lies about her ex and general behaviours. Ignoring him for days again, threatening to game quit if she doesn't get the attention she wants. It'd all gotten so tiring that he didn't have any attraction to her anymore. He had no sex drive. He dreaded seeing her. But had to constantly message her. He's been feeling like this for months. Karl didn't want to leave her just before her birthday, he felt it cruel. But then it was the anniversary coming up. He didn't want to be responsible. He'd tried gifts, trips, anything just to make her happy. No matter what he did he still felt like nothing was enough. I managed to talk him through. About threatening suicide if someone wants to leave, is indeed abuse. He wasn't himself. How we felt and how we where worried. He got choked up. Not realizing anyone cared. He asked if he should leave. I asked if he was happy. "I can't imagine not having her there." Okay no. Not what I asked. Eventually he confessed He'd never felt lower. I said. Can you see yourself marrying her? No. Infact he said the thourght freaked him out. I said. Well. Why are you with her. Eventually it got to a point He left her. She said she'd been thinking about it. Yay? No 12 hours later he calls me saying its all fixed. Its all okay. How He's a horrible person for doing this to her. How it's him that needs to change. How he will spend a long time making this up to her. You know. I'm a domestic abuse survivor. But I never realized how much hearing that killed abit of my sould. Trying to convince Karl that he's worth anything is like trying to convince a deaf non signing American Conservative that the gays aren't trying to make him gay too.
They do eventually a few months later split. She says she wants to breakup as he "doesn't love her the way she wants him to" he is hurt but says okay. She then obviously realizes hey, he isn't gonna start begging on his knees. You can only hurt someone so much. She then asks "breakup sex" directly after and to this day its our favourite quote. But he says no, she asks for one more night, he says no you just broke up with me? Leave? She complains about not being able to get to the train station. Now. Karl didn't have his licence till a few weeks later. So queue the weirdest car ride with his DAD you've ever heard of. She cried. Hugged him. Begged him to reconsider. Karl officially has realized how disconnected he's become. Nah.
Queue a weird amount of messages ranging between "I'm sorry baby" to "I CANT BELIEVE YOURE GIVING UP ON US" and sexually charged messages, After karl finally blocks her. She begins to call him from various different numbers. Tries to get with his friends. Fails. Still calls him crying for the next 6 months. In which these events happen.
Frank from before. Now it turns out. While we don't have full timeline but either weirdly around the time they broke up they got /very/ close. To the point despite Frank having a partner. She was begging him to sleep with her. But Being weird with it. One minute she wants him. Next she doesn't. Basically, she loves the idea that she could have him. But doesn't want to keep any of them. Frank had a girlfriend. Goddess of a lady. Daisy. Regina proceeded to pick on every little thing to Frank about daisy she could. Always. Physical appearance.
Then. Now I am simply not making this up. after Frank separates himself from this situation. Regina begins to harrass Daisy, With telling Daisy about how much Frank's missing out on not shagging her instead.
And making 6 different instagram accounts to harrass them, and this is where I come in further.
Regina now, after the hate group chooses some last straws she can pull to drag him back. She makes a fake account. Goes to message Frank. With the opener of gossip about me and my partners sex life. I talked to Regina less times than I can Count on one hand.
The main one I'm aware of is "Did you know my partner drinks my names piss" Which I'm not here to kink shame; but this does not happen unfortunately but i still find it beautiful of a statement.
I one day due to some more harassment and more attention than I'm used to.
Decide to private my instagram. It was only for 24 hours in full so I could change some settings and archive some things. Within 15. An account. David, requests to follow. Strange. Cause my account is shadow banned and cannot be shown to non followers. I click. Heavens foretold dear friends. Regina's new boo. Id like to clarify. 2 weeks before Karl was still getting snotty teary calls telling him she misses him. Karl's friends where sending screenshots of Regina trying it on with them then getting snotty when she was rightfully laughed at.
I ask "hi??"
"Hi me and my girlfriend just wanted to stalk how cringe your posts are" I wish I could have been funny and not caught off guard. And shamed them. Oh god. I wish I had. Basically I told him, the gym is waiting. She will chew you up. Idk what I did but I'm sorry. Godamn. Leave me be. And they said "It's not that deep lighten up" I am indeed embarrassed.
But they kept mentioning my workplace. I am a bartender, and one day she did come in with a man, they seems very loved up but then again. It certainly wasn't this guy. then said bad things about me infront of a coworker. It was a little satisfying seeing her face fall and hit the table from shame as I was carrying an ice bucket past her. She was already cut off at this point for her antics.
David's best friends memepage now follows me. But has been the first out of 5 accounts not to say anything. I'm sure they think I don't know. David claimed I was lying in my encounter. I do wonder if I could flip the table entirely.
but I also wonder if she's just very mentally unwell. But it's been 1.5 years of this and I'm just abit knackered and pissed off.
I'm 20 feeling like a highschooler. But I'm working for a bipolar diagnosis and I have ADHD, the paranoia. Is driving me up the wall man. Like this woman knows enough of my details and she's spread where I work. She's been to my house. She has clearly gotten multiple people involves historically and despite me trying to apologise, it makes no difference.
If I knew what the issue was, I'd gauge it. But it's not knowing and not being told. But it's reassuring it's not just me. With daisy, I'm wondering if this is historic. Might be vanity? She (used to?) Post alot of ...suggestive photography and always wears a lingerie corset and heavy makeup, filters. Nothing wrong with that of course but she's a very sexually orientated person, and given the contexts to that behaviour. I wonder if its to cover some in depth issues. But that's just a theory. Part of feels hey, if she needs men to tell her that I am ugly, cringe and worth nothing. Then she van have that. The other half makes me want vengeance for the boy, prove that I'm not whatever she'd been making me out to be and make her realize she needs to change. But that's. abit pathetic innit.
Anyway I doubt anyones made it this far and if you have. Thank you for reading my story and the weirdness of it. I hope it hasn't been too shit. Just needed to get it off my chest. And maybe if anyone has anything to say.
TLDR: my best friends ex has always had an issue despite my efforts. Getting various people to harrass and bully me, She tried to get with his friends, other guys we knew and harrassing us all. All while still crying she misses him. Her new bf thinks I'm lying and is joining in, his best friend now follows me too. My partner allegedly drinks my piss <3

If you've never played League of Legends (abbrev. to LoL) before, there's just a few things you need to know:

1. It's a 5v5 team game
2. The playerbase is notoriously TOXIC as hell
3. 'Faker' is the best LoL player Esports has ever seen, and even though he's played professionally for 10 years, he continues to get better and better
4. A 'Skin' is an alternate appearance for a Champion (playable character) that you can buy with real money (but first you need to buy ingame currency called RP), or possibly get for free through freebie rewards from playing the game
5. The average price of a Skin is $10 USD (1350 RP)
6. Faker does not equip his Champion with a skin, opting to only use the default (free) appearance

Things begin 6 days ago when Riot Games posts a teaser trailer for a documentary film about Faker, titled "Hall of Legends: Faker', and it's set to release on the League of Legends Youtube channel June 14th. As part of the hype, 4 days ago, another Hall of Legends teaser video was released, this time showing a brand new Skin for the Champion Ahri, titled, 'Immortalized Legend Ahri'. It should be known that Ahri is one of the most popular Champs in the game, and Riot Games has continued to capitalize on that fact by producing more and more Skins for her.
Before this announcement, Ahri had 17 skins. Now, she will have 20, because this new Immortalized Legend Ahri skin has 3 variations, priced as 3 different bundles. Yes, this means, if you want any of these Skins, you have to buy one of the bundles that come with a bunch of other stuff, oh and the Skin too. Riot has also announced that 30% of the proceeds of this Skin line will go to Faker. So here are the prices:

Risen Legend Ahri Bundle (Basic): 5430 RP (~$40 USD) Immortalized Legend Ahri Bundle (Fancier): 32,430 RP (~$240 USD) Signature Ahri Bundle (Amazing!!!): 59,260 RP (~$500 USD)

How does leagueoflegends react? Let's take a look at the top thread as of this post: (Commenters will be C1, C2, C3, etc.)

C1: The way i thought 59k rp was a typo lmfaooo C2: i had to go look, then check the store, then go look again, then check the store again, and im still not completely sure this is actually real. i hope zero people buy the skins C3: My first thought was they wanted me to tell us we would get 59.620 RP of "value" from buying the battle pass C4: $500? Yeah no thanks lmao. Faker doesnt even use skins anyway. C5: If you wanna be like faker, don't buy a skin 💯 C6: For that kind of money I almost expect Faker himself to deliver me a code to the bundle C7: I would expect him to come home and play the next 10 matches for me and also go in a date together

One Redditor points out the irony of having the most expensive Skin in the game in honor of Faker, who doesn't use Skins:

C1: "Let's celebrate the face of our e-sport (who famously doesn't even spend money on skins) by absolutely overcharging the fuck out of his event skins." What the hell Riot lol C2: how ironic is that Faker himself is against using skins and they give him the most expensive skin ever made lol C3: Why is he against using skins? C2: He said before he sees skins as a waste of money and that the base skins are good enough for him, he has numerous world skins under his name but he used them less than 5 times in a 11 year career C3: The game is free, how else would it generate income if not for cosmetics? In its early days at least, I have no clue if the professional scene generates enough income to make it sustainable. [gets downvoted] C4: The comment above is straight up false information. Faker growing poor and couldn't afford skin and it became a habit of him so even with alot of money now he still don't buy skin. C3: What does he even spend money on? Doesn't he play league 24/7? [more downvotes] C5: Food, housing, and his future. Why would he spend money on a free game when he can save it and be 100% financially stable for the rest of his life. It's pretty pointless to spend money on league. C6: Pretty sure [T1, the team he plays for] covers his food and housing already lmao C3: [to C5] I'm sure that faker buying skins here and there would bankrupt him. C5: You're missing the point. People who grow up in poverty are generally going to avoid spending on things like that. It's not that it's going to bankrupt him, you're twisting my words in bad faith.

Several Redditors do the math to help themselves visualize the price of the most expensive bundle:

C1: $500 for a skin, that's a real signature move by Riot. C2: "most generous bundles" (40$ 💀) C3: Elden Ring´s DLC is cheaper C4: This is giving me real "pride and accomplishment" vibes lol C5: Would you rather have a) A couple of skins and some useless bs like icons and emotes b) A Steam Deck with a dozen of the best indie games ever made C6: c) rent money C7: Almost half my monthly wage 💀💀 C8: 3rd world country mfs got hit harrrd This shit is 2.5x my monthly 9 to 5 wage And I mean the full thing, no rent no food no bills nothing 2 and a half months of ice soup if I wanted to buy this lmao C9: This is like, 10 months worth of rent for me what the fuck C10: This is what I earn in a month (working 10h) 💀 I can decide between food and league skin

The comments keep coming in, so get popcorn ready, because there's more threads to read after that!
The Hall of Legends event was never about celebrating Faker, it was to use him to make money.

Almost every League player loves Faker and we would have loved to to own the Ahri skin dedicated to him. Riot putting the full skin behind 600$ paywall is the biggest insult imaginable to us. Riot chose to exploit our love for Faker for their own gain.

Meanwhile, in AhriMains, the Fox lovers are alternating between cope memes and boycott ideas:
I Blame the Whales for this.
Easy way to counter Riot's prices
The best way to rebel is to boycott Ahri herself

I think it’s atrocious that riot is pricing a skin bundle that costs half my paycheck for an entire month. I say let’s protest in a way that forces riot to the negotiating table. All we need to do is ban Ahri every single game. The whales will never get to enjoy the skin if they cannot play the champion, and then riot will be forced to lower prices. Similar strategies worked for ryze. Let’s work to get Ahri banrate to 100%. Edit: Big brain idea, link this post in every pre-game lobby and ask them to also do the same in their next game. Exponential spread. Also send this message below to all your friends and ask them to do the same, without the parenthesis. Grassroot movements need commitment. "Discourage riot from 500$ skin. Send to friends. https://old.reddit.com/AhriMains/comments/1d2rsi5/the_best_way_to_rebel_is_to_boycott_ahri_herself/?sort=top"

Reminder to not piss in the popcorn.

Hello everyone! 🌟

I was watching the Pokimane episode, and she said something that really hit home about the need for online security information, especially for women. She mentioned how scary it is to deal with stalkers, tracking devices like AirTags, and the lack of adequate stalking regulations. This really hit home that we need to be vigilant.
I've had people try to hack my accounts and make fake profiles using my photos. It's a bit much. The law hasn't quite caught up yet, so we need to stay vigilant! Here are some tips to keep us all safe online:
This will be a long guide but it's so important to stay informed.

🌐 General Online Safety Tips

Use Strong Passwords:

• Create unique passwords for each account: Use a mix of uppercase letters, lowercase letters, numbers, and symbols.
• Avoid using easily guessable info: Steer clear of using birthdays, pet names, or other obvious choices.
• Consider using a password manager: I use 1Password, and it’s a lifesaver! A password manager securely stores and generates strong passwords for you. It was a bit of a hassle to set up, but it’s so worth it now because it helps me store everything. Plus, I can share my Netflix password and other accounts with my family easily. 🌸
• Need help setting it up? They also have great user guides here and support here if you get stuck.

Enable Two-Factor Authentication (2FA):

• Extra layer of security: 2FA requires a second form of verification, like a text message or app notification, in addition to your password. It’s like having a double lock on your door, making it much harder for anyone to break in! 🔒✨
• Why it works: Even if someone gets your password, they’d still need your phone or access to the second verification step to get into your account.
• Always enable 2FA: Turn it on for accounts that offer it, especially for your email, banking, and social media. It gives your accounts an extra level of protection. 🌟
• Need help setting it up? For more details on how 2FA works and how to set it up, check out this detailed guide and JumpCloud's explanation.

Be Cautious with Personal Information:

• Avoid sharing personal details: Don't post your address, phone number, or financial information online. This info can be used by scammers or stalkers, so keep it private.
• Be mindful of social media sharing: Think twice before sharing personal updates. Make sure you know who can see your posts and adjust your privacy settings to keep things tight.
• Learn more: For additional tips on staying safe online, check out these great resources from Stay Safe Online.

📧 Email Security

Beware of Phishing Scams:

• Be cautious of unknown senders: If you get an email asking for personal info from someone you don’t know, be super sceptical. Scammers love to pretend they’re legit to trick us.
• Watch out for suspicious links and attachments: Even if it looks like it's from a friend, always double-check before clicking on anything. Phishing emails can be sneaky!
• Verify the sender: Check the email address carefully and look for signs of phishing, like bad grammar or urgent requests that seem a bit off.
• Learn more about phishing: Dive deeper into spotting phishing scams with Phishing.org.

Use Secure Email Providers:

• Choose providers with strong security measures: Popular email services like Gmail offer robust security features such as two-factor authentication, spam filtering, and protection against phishing attacks. These are great for general use and offer strong security.
• End-to-end encryption: For the highest level of privacy, consider using email providers that offer end-to-end encryption, like ProtonMail. This ensures that only you and the recipient can read the message, and not even the email provider can access it. This is particularly important if you're handling sensitive information or just want the peace of mind that your messages are truly private.
• Robust privacy policies: Ensure your email provider has a solid privacy policy that protects your personal information and doesn’t share it with third parties. Gmail has strong security, but it’s important to understand how your data is used by reading their privacy policy. ProtonMail, on the other hand, is designed with privacy as a priority and doesn’t collect or share user data.

🔒 Social Media Safety

Adjust Privacy Settings:

• Review and update regularly: Regularly check and update your privacy settings to control who can see your posts and personal information. This ensures you're always in control of your online presence.
• Limit visibility: Set your profile visibility to friends only or create a custom list of trusted people. This helps keep your personal information and posts private.

Limit Location Sharing:

• Avoid real-time sharing: For safety, avoid sharing your location in real-time. If you want to share your location, do it after you've left the place. This prevents anyone from knowing your exact whereabouts at any given moment.
• Be mindful of tagging: Be cautious about tagging your location in posts or stories. This also applies to tagging friends; make sure they're okay with sharing their location too.

Be Selective with Friend Requests:

• Accept only known contacts: Only accept friend requests from people you know in real life. This helps prevent scammers and stalkers from accessing your personal information.
• Be wary of strangers: Be cautious of friend requests from unknown people, as they often create fake profiles to gather personal information. Look out for red flags like a lack of mutual friends or minimal profile information.

🛡️ Device Security

Keeping our devices secure is just as important as online safety. Here’s how to make sure your gadgets are protected:
Keep Software Updated:

• Regular updates: Always update your devices and applications to fix security vulnerabilities. This includes your phone, computer, and any apps you use regularly.
• Enable automatic updates: Turn on automatic updates so you don’t have to worry about missing any important security patches.
• Learn more: Check out this information from US-CERT.

Install Antivirus Software:

• Use reputable antivirus software: Protect your devices from malware and other threats with reliable antivirus software. Some popular options include Norton, McAfee, and Bitdefender.
• Regular scans: Schedule regular scans and keep your antivirus software updated to ensure maximum protection.
• Read reviews: Find out which antivirus software is best for you by reading reviews from AV-TEST.

Use a VPN:

• Encrypt your connection: A Virtual Private Network (VPN) encrypts your internet connection, adding an extra layer of security, especially on public Wi-Fi networks.
• Protect your data: VPNs help protect your data from being intercepted by hackers. Some popular VPN providers include NordVPN, ExpressVPN, and CyberGhost.
• Learn more: Discover more about VPNs from NordVPN.

💬 Communication Safety

Use Encrypted Messaging Apps:

• Choose apps with end-to-end encryption: For private conversations, use apps like Signal or WhatsApp, which offer end-to-end encryption to ensure that only you and the recipient can read the messages.

Be Wary of Strangers:

• Caution with unknown contacts: If someone you don’t know tries to contact you, be cautious. Don’t share personal information or agree to meet them in person without verifying their identity.
• Learn more: Get tips from Get Safe Online.

💖 Mental Health and Safety

Seek Support:

• Reach out for help: If you’re experiencing online harassment or threats, seek help from trusted friends, family, or support organisations. Every situation is different, so it’s important to reach out to the appropriate authorities or professionals who can provide guidance specific to your circumstances.

Remember, your safety is paramount. I could yap all day about this, but for all of us (the girls, the gays, the theys) who are more susceptible to harm, the law doesn't usually catch up until tragedy strikes. So let's stay vigilant! 🌟💖

I’ve really enjoyed using my three CPTs with boosters and display to smoke briskets and pork butts, roast chickens and roast the Thanksgiving turkey. It’s an incredibly useful, well-designed and well-built family of products.
Except for two things: lack of charge indication and reliability of MeatNet.
Chris has explained that a special chip would be required in order for the CPT to report its charge state, and it isn’t compatible with high heat environments. I can live with that. So far, I haven’t had an instance of a CPT running out of charge during long cooks going well over 12 hours. But when I asked why the booster doesn’t have charge indication, Chris said it just didn’t make the cut for the first generation of hardware. OK, but this weekend my overnight Memorial Day cook was nearly ruined due to lack of information about booster charge, and maybe a problem with MeatNet.
Here’s what happened:
I smoked a brisket point and flat, separated, with CPTs in each cut. Before the cook, I pressed the button on each CPT’s booster and got a green light, so I knew the probes were fully charged. The meat went in my Kamado Joe ceramic grill, and I put a booster on a cabinet about six feet away. I put the other booster on a table indoors about 40 feet away. I took the display and my iPhone to the bedroom, about 40 feet from the second booster. I set the iPhone to Never Lock and brought up the CPT app so I wouldn’t miss the prediction alarms (learned to do that the hard way.) I set the CPT predictions to 180F to catch the end of the stall, at which time I’d think about wrapping. One of the CPTs set off an alarm at 180F. I decided not to wrap, and reset the predictions on both CPTs to 200F, at which time I‘d wait until they hit 203F and remove the cuts for a long hold.
The alarms never went off. Luckily, I happened to awake in time to see in the app that both cuts had reached and slightly exceeded the target temp. But both CPTs were not connected, and the display was showing they weren’t connected. I rushed to the smoker with a ThermoPen and was relieved to find that the two cuts weren’t far over the target temp of 203. One was at 205F and the other was at 208F. The texture of the cuts was good (wabba wabba) and a ThermoPen showed temps weren’t too far over target over most of the meat. I took the meat off for a long hold.
What happened? It didn’t take long to discover that both boosters had run out of charge. Given the last temps shown by the app, I guess I was lucky to awake shortly after the boosters ran out of charge. I’m puzzled why display and/or iPhone didn’t directly connect to the CPTs in the ceramic cooker when the boosters ran out of charge. The total distance was less than 75 feet. Is it necessary to have a MeatNet relay close to the CPTs? I’ve been doing that, but I thought the range was better. Is the distance significantly reduced when the CPTs are inside a ceramic grill? If this is the case, I think I should put the display, which has a much larger battery, near the ceramic grill so the iPhone can stay connected anywhere in the house.
So, why did the boosters run out of charge? The last time I can be sure I charged the boosters was about month before, after smoking a pork butt. I did one cook after that, a roast chicken last week. At the time, I pressed the buttons on all three boosters. Two lit up green, and one didn’t light up at all. I put the CPT in one of the other boosters and it showed green, so it looked like the CPT was charged but, unlike the other two boosters, its booster had run out of charge. Seemed like that shouldn’t have happened in 2-3 weeks without using the CPT. In fact, I had seen this once before, but I didn’t mark the booster that had run out of charge earlier than expected. So I charged up the one that had run out and left it on my desk with a different CPT to see if it would happen again (I wanted to rule out a problem with the CPT.) When it came time to do the brisket cook on Memorial Day, the booster in question showed green when I pushed the button, which meant the CPT was still charged. Of course, that didn’t tell me what state the booster battery was in, but I had to get on with the cook so I aborted that experiment. I thought I had charged the boosters after the roast chicken cook, but maybe it was only the booster I was testing.
It’s possible I have one booster that can’t hold a charge, but it’s also possible that all three of my boosters have problems holding charge. It’s also possible I have one or more CPTs that draw power from the booster too often. After all, my CPTs and my boosters are early production models. I need to do some careful experiments to figure this in out, and the lack of any charge indication for the CPTs and boosters is going to make that very difficult. One idea I had is to use a USB dongle I have that reports mAh delivered. That might give me some info by showing how much charge a given booster can take.
I’d appreciate any advice for setting up experiments to determine if I have bad boosters and/or CPTs.
Beyond that, I hope Combustion will consider a Gen 2 booster with charge indication (fine if it’s in the app.) I will probably replace my boosters if that happens (though a trade-in deal for early adopters would be nice.)
Recently, someone posted about Combustion offering a charging dock for CPTs. After this weekend, that idea appeals to me, too. The ability of the boosters to keep the CPTs charged would be superfluous (the plain charger would do), but it would still be worthwhile for keeping the boosters charged for use in MeatNet.
The other problem I’ve had is frequent disconnects. It’s worst in my Breville Combi Wave oven, but I’ve chalked that up to the amount of metal between the CPTs and boosters, and maybe noise generated by the Combi Wave electronics. The CPTs do reconnect eventually, but sometimes it can take several minutes. I can get through a roast chicken cook (oven, not microwave), but the disconnects are annoying. Most of the time I don’t get lots of disconnects when using CPTs in the ceramic cooker. They can go hours without disconnecting. But I do get them. The CPTs eventually reconnect, but it‘s annoying. Perhaps this happened as the boosters charges were winding down to zero.
Oh, one more issue: After setting a prediction, sometimes I couldn‘t get back in to change clear the prediction. When I tapped the prediction section, it would display the food safe page. Then, sometime later, I could tap the section and change or clear the prediction.

Back in the day, when Google Maps and Google reviews first became a thing, I relied on them quite a bit. It was an easy way to get an idea of how a particular product or service was doing.
Over the last few years however, I’ve increasingly noticed how useless the platform has become.
Whenever I do research on dentists or physiotherapists for example, nearly all of them have hundreds of only 5 stars, with commentary so generic it may as well have been generated by A.I.
None of it is believable anymore and when you do want to leave honest, constructive feedback, it either gets blocked from public view or removed by businesses trying to keep their score artificially high.
On the other side of the coin, businesses that are legitimate can end up receiving unwarranted, fake or bot reviews that Google fails to remove, despite going through the usual channels to resolve it.
Just seems to me that a once useful tool is now failing everyone and reminds me of the crazy world of censorship that we all live in.

I’ve never really posted on a forum but I just need to know if anyone else relates or has any type of insight/advice here. I will say that I have OCD, am AFAB, pansexual (20s) and I am set to see a trauma psychologist very soon.
I was very adamant with family about being seen as boys from some show but it was on and off. I remember having a crush on them but I can’t understand why I also wanted to be them. When I hit puberty I started pretending to be a guy where I felt safe enough to. I was okay to be a girl sometimes I feel like despite being unhappy with insecurities, but I had this inexplicable confidence when I was presenting myself as a guy. I am prone to both maladaptive daydreaming and dissociation, and I’ve been doing it all my life.
I had an upbringing that was extremely toxic so there’s no telling what has been pushed onto me. I have felt like a lot of the years of imagining myself as a male I both wanted to be and to be with was me trying to nurture some father wound, because I would imagine myself as a male nurturing the female version of myself, I just hardly knew her. I’ve rarely dissociated as a woman, but there are countless times where I’ll compare myself to women to see if I’m desirable in stereotypical ways.
I dated girls when I was younger and presented as a male, but part of me felt like I was so obsessed with being this cool, attractive guy that I put up a front. I felt I had to be tough and a certain way, so when I got so tired of being that, my feelings of masculinity would crumble and there was that vulnerable girl.
I chose to go by they/them, telling myself that I only wanted to be masculine presenting because I didn’t feel attractive as a woman. But somehow this desire keeps coming back. I see I have some gender dysphoria, but I keep feeling like I’m just faking everything and putting on a mask to feel better about myself.
I see my chest when I look masculine in the face and hair and I just kind of sigh but I’m not in distress. I think I would like to feel/do many things as a man, but I’m scared I’m mistaking this for something else underneath. If I could’ve been born a man I would be glad because then I’d have it all already and I guess maybe I’d be content, I don’t know. I know I feel disappointment and sadness with the idea that I possibly don’t really want to be one.
TLDR; I keep getting confused on whether me dissociating as a male character through my life is only because I find them attractive or I truly want to be male/masculine presenting. I feel different because it doesn’t always feel like my natural self, I just think I wish it was but I sometimes feel like I’m forcing.

When I log in and open my tree, it is not opening to the last tree I worked on, which is generally mine, but to a mock tree I rarely if ever look at, I so have to leave this accidental choice it's routing to, go to my tree list and find my tree among the offerings wait for it all to re lode and open it that way.
When I was in the Ancestry Beta test group nightmare, every time I worked on someone and was going back to my hit results for them it sent me to another person in the tree, like their parents, rather then them and or to this tree. I suspect that I did something dumb like made myself the tree holder. That function is always glitchy for me. Often it wouldn't take the tree I'm adding, or it will take it, but then I can't find it, listed with the other to delete tree drop down list.
Does anyone know how to get it to route to my tree as the continual fail safe when I am using it? In the past the fail safe was that it routed to the last tree you added something to, or if you rewrote the overview. So you added a period to over view or corrected a typo, or added a fake individual and it fixed it. None of that has worked.
My tree is set to private unsearchable, can I even link it to me and still get my matches? I don't get Thrulines, but have never had this issue before. The tree won't be there or the individuals I addded, but never had a tree i was not working on randomly show up each time I open my tree.
My hubby fiddled and got me the Old Ancestry display I like, so now not being tortured by new Beta, but this appears to be it's only troubling legacy of annoyance and it's still routing me to the incorrect tree, just as it was under new Beta, and not to my DNA but to one of m oldest mock trees.