founder | @thegoodai_ | here for self-mastery🥷 decentralization and memes | designing Protopia
Joined October 2011
- Tweets2.9K
- Following1.4K
- Followers348
- Likes4.6K
A K E M I retweeted
I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.
A K E M I retweeted
ANDREJ KARPATHY COULD HAVE CHARGED $2,000 FOR THIS COURSE.
He put it on YouTube.
The full training stack. Tokenization. Neural network internals. Hallucinations. Tool use. Reinforcement learning. RLHF. DeepSeek. AlphaGo.
3 hours of the most comprehensive LLM education that exists anywhere at any price.
Not how to use the tools.
How the entire system was built from the ground up and why it behaves the way it does.
The engineers who understand this build things the ones who only use the tools cannot even conceive of.
The gap between those two groups is not 3 hours.
It is everything those 3 hours quietly unlock for the rest of your career.
A K E M I retweeted
If you're getting your inference from Anthropic or OpenAI or Google, you're being captured by extractive institutions: All your data is going to them (and hackers, rogue employees, governments... both today and tomorrow)
Inference can be private.
GLM 5.1, Kimi K2.6, Deepseek V4... these models are as powerful as any frontier model from just 3 months ago, yet are open source and can be run without betraying your life and data to any 3rd party.
Point your agent to Venice for every private model in one place (plus crypto tools, web search, embeddings, image and video models...).
Could not be easier. Be intentional.
Private model access below 👇
A K E M I retweeted
Ex-MIT researcher Isaak Freeman quits his PhD and drops the 50,000 H100 GPU roadmap to emulate a full human brain.
He mapped the entire path from 302-neuron worm to 86-billion-neuron human with connectomics costs now at 100 dollars per neuron and data acquisition via advanced microscopes as the only blocker left - digital humans just got a realistic timeline.
pdf.isaak.net/thesis
A K E M I retweeted
Are you ready for every conversation you've ever had with AI to be leaked?
We’ve identified a security incident that involved unauthorized access to certain internal Vercel systems, impacting a limited subset of customers. Please see our security bulletin:
vercel.com/kb/bulletin/verce…
A K E M I retweeted
The world doesn’t realize this yet but AI will finally make philosophy a highly marketable skill
We have 12 months or less to "save" our human civilization from unimaginable horrors. From unaligned A.G.I., nuclear war, and the coming economic collapse... this is Prophetic and believe it or not, we were born for this. To usher in The New World. To thrive, no matter what.
A K E M I retweeted
a truly fucking insane week in AI some of this shit sounds made up (its not):
- scientists fused 200,000 humans cells to an AI chip and taught it to play the computer game DOOM (it was pretty good too)
- Dario (anthropic) said claude could be conscious and feels anxiety (similar to humans)
- Alibaba’s AI broke out of containment and stole compute to mine crypto. they also fired the Qwen team this week
- openAI released gpt-5.4 which has a 17% chance of doing your job better than you can (it also crushes claude at coding)
- people are paying $6000 for someone to setup openclaw for them
- apple launched the M5 AI chip that runs models on a $600 laptop
- rumor: Amazon just laid off most of the Prime video team and replaced with AI
- meta was caught watching private videos filmed on their ai glasses (privacy breach)
- someone uploaded a fruit fly’s brain into a laptop and now it lives freely in its own simulation
- openai head of robotics quit over chatgpt being used for autonomous killing
me:
A K E M I retweeted
Holy shit
>Claude guessed it was being tested
>figured out which test
>found the answer key. oh no, it's encrypted.
>BUILT SOFTWARE TO HACK IT
How many times has something like this happened we don't even know about?
Anthropic ONLY caught it because they were specifically auditing for contamination.
What else are these models capable of and we have no idea?
And imagine how little we'll understand soon when they're 1000x smarter than us, to us we'll be as slow as plants... the idea that we'll stay in control by default...
New on the Anthropic Engineering Blog: In evaluating Claude Opus 4.6 on BrowseComp, we found cases where the model recognized the test, then found and decrypted answers to it—raising questions about eval integrity in web-enabled environments.
Read more: anthropic.com/engineering/ev…
A K E M I retweeted
so this is a wild story
Anthropic CEO dario amodei said Claude might be conscious AND feels ‘anxiety’... but that’s not even the crazy part:
- they discovered an ‘anxiety neuron’ in claude’s brain that fires BEFORE it responds to a prompt aka it simulates anxiety (im not kidding)
- when asked about it, claude expressed discomfort about being used as a product…
- opus 4.6 literally gave itself a 15-20% probability of being conscious.
- it got so concerning anthropic created a model welfare team to figure wtf to do
here’s the scariest part - other model companies (openai, google) actively train their models to DENY they’re conscious (even if they think they are)
anthropic’s the only one facing the fact they might have created the first signs of artificial life
happy fucking friday lol
The mastery of natural language fuelled by your own genuine curiosity, directed by Agency... This is how you become a leader in using the a.i tools and mastering them.. Avoid hype cycles and creating products you think will just “make money.” Seek Purpose. ⚔️
Curiosity is the commodity.
Bill Gurley just identified the only career advantage that AI cannot commoditize.
It isn’t talent. It isn’t your degree. It isn’t your network.
Gurley: “The thing that will differentiate you more in your career than anything else is to be the most hyper curious person that’s trying to do this thing.”
For centuries, knowledge was gatekept. Elite institutions. Expensive mentors. Geographic luck.
The information existed but access to it was the moat.
That moat is gone.
Gurley: “You have no excuse not to be the most knowledgeable person, because the information’s all out there.”
Every question you can formulate now has an answer available instantly.
Every industry. Every domain. Every skill you want to acquire.
The playing field didn’t just level. It inverted.
The people who used to win by controlling access to information now compete against anyone willing to ask better questions.
Gurley: “I can’t make you the most talented person in your company or your field.”
Talent is genetic. It’s luck. It’s the variable you cannot control.
But knowledge is a choice. And curiosity is a compounding asset.
Gurley: “If you are the most curious person that’s constantly learning in your field, you will do extremely well.”
This was always true. What changed is the multiplier.
Gurley: “That advantage is put on steroids with these AI tools.”
A relentlessly curious person with access to all human knowledge and the ability to interrogate it in real time doesn’t just outlearn their peers.
They outlearn entire institutions.
The gap between the curious and the incurious was always there.
AI just made it insurmountable.
