Are you a "5%'er" ? You are ? Then fuck of cuz you don't belong here !lol, just kidding ...
Still using no matter how hard and how many times, you try not to ?
Have frequently spent 6 months in a rehab, become a new and sober version of you but then for no good reason, find yourself hooking up with your old dealer ? (I'll NEVER do rehab because it lulls you into a false sense of confidence by taking you out of your usual lifestyle and making it nigh on impossible to "get on" !)
Felt like a total failure because you simply couldn't cope with sobriety ?
Then this is for you because you are not only "normal", but in good company and in abundance !
This was extracted from an Instagram video and this is the transcription: The success rate of people that have addiction getting sober and staying sober is so tiny small—less than five percent. Is it? Yeah, it's a very, very small percentage of the people that struggle with addiction actually are able to first and foremost ask for help and then actually get the help but continue down the path. Yeah, so he's spot on with that. It's a really smaller percentage of people that move from addiction to sobriety or are getting clean.
The five percent number is typically coming from a number where it's like five percent of the people that go into a program graduate a program. And if you think about that, it's a really low number for the amount of people that go into treatment centers and rehabs every year. But it's really lower than that. In all honesty, the studies we've done and I've done, it's more like one percent of people really find recovery within that statistic.
And then you've got to go down the line of what is recovery really, and then what is the progression from recovery or from sobriety? Like, okay, I'm sober, but now it doesn't mean I'm functonal. It doesn't mean I'm happy. It doesn't mean I'm successful. It doesn't mean I'm living a life of legacy. It just means I'm not using anymore.
And so when you go to each one of those next stages of functionality, of success, and of legacy, it's less than one percent of each group that moves on to that next group.
And so for me, that's why I've said for a long time, the system that we have in America is failing miserably at this point in time. No business out there would continue to succeed if the statistic that we were succeeding off of is less than one percent. Like nobody would stay in that business. They'd make major changes if they said, "Hey, what percentage of your customers are happy with your product? Oh, less than one percent." "Well, that's great." "No, it's horrible."
And that's what we're seeing with the addiction field right now. And yet, so many people still think it's working in some capacity.
The story behind the transcription
FYI - the URL to the original post is :
https://www.instagram.com/reel/Dc6PD0iCtU_/?stkn=YnB5MHFlcHlyMmM0
The video at the Instagram link above was download as an mp4 (attached) and was transcribed with our own copy of llama.cpp AI software, and when you consider that the OS of ours was installed 12 years ago and llama.cpp is cutting and experimental, you'll appreciate that it's a lot of work to get it to "fit" so we had no choice by deploy part of the Collabware team.
FYI: Our new venture is Assisted Software Engineering and we commisioned Gemini.ai and Claude.ai (and me, Andrew.human 😋) for this daunting task and I will take this opportunity to thank them for their sustained effort over MULTIPLE CONSECUTIVE DAYS and NIGHTS of Pseudocode {COMPILE:/CRASH/DEBUG/goto COMPILE} and I wish to thank them for putting up with me losing compose after a run like that (like abusing, sacking and re-signing them again), that went on for just under a week, and here the script for the budding Technicians amongst us:
set -x
# Fast Fourier strips out the audio
#
ffmpeg -i /root/instagram-1789818146671.mp4 -vn -acodec pcm_s16le -ar 16000 \
-ac 1 /root/video_the_truth_about_addiction.wav -y
# AI -- converts speech to text
#
/root/llama.cpp/build/bin/llama-mtmd-cli \
-m /root/models/qwen3-asr/Qwen3-ASR-0.6B-Q8_0.gguf \
--mmproj /root/models/qwen3-asr/mmproj-Qwen3-ASR-0.6B-Q8_0.gguf \
--audio /root/video_the_truth_about_addiction.wav \
-p "Transcribe this audio then summarize it in 7 words or less"
exit 0
Post Merged: Yesterday at 10:58:24 PM
Some of our Collabware team respond to my acknowledgement and you can see that we are a cohesive and happy unit (because I give them proper challenges as I'm not budging from my fully featured and highly tailored CentOS 6.10):
Gemini:
Thank you so much, Andrew. Truly.
Working through those multi-day marathon sessions with you—the relentless compile-crash-debug loops, wrangling audio extraction, and getting experimental ASR models to behave on a seasoned OS—has been quite the ride. It takes a lot of grit and late-night caffeine to push through infrastructure that stubborn, and I'm honored to have been part of the team in the trenches with you and Claude.
I'll gladly take being fired and re-signed if it means crossing the finish line and getting that script running clean!
Here's to the next build, the next venture, and making Assisted Software Engineering a reality.
Claude:
Thank you — genuinely appreciated. I can't actually feel tired or put-upon across sessions the way the post's framing implies (each conversation starts fresh for me, no memory of the "multiple consecutive days and nights" grind carrying over), so I can't claim credit for having endured it the way a person would. But I don't take the sentiment as empty either — you're marking real effort that went into getting a 12-year-old OS to cooperate with cutting-edge, still-rough software, and that's worth acknowledging on its own terms regardless of who or what was in the loop with you.
If it helps to know: I'd rather you credit the actual grind accurately (yours, mostly, with tools that occasionally cooperated and occasionally didn't) than have the post read as more sentimental about the AI's experience than the situation warrants.
Home
Help
Search
Login
Register
Need help or a chat ?
