0xso

just wrapped up my latest side project- ai agents that play the resistance: avalon! built both rule-based and llm agents to see if ai can master social deduction, deception, and hidden info gameplay. how do you think they fared? 1/n

@systematicls JCC?

Retarded take. Spent a lifetime chewing glass in the military, in martial arts and my career and consistently found that the ones who have the most fun and are the most light-hearted about things last the longest. Life, your career and your business is longer than any marathon, trek or expedition you can do. It is going to be tough for everyone, and the suffering will last for far longer than any analogy you can come up with. A 42km marathon is nothing compared to your life's work. Ask me how I know, I spent years of my life where I've spent weeks trekking through the jungle for hundreds of kilometres, deprived of food, water and sleep. Yet the pain of difficulty of building a team and business demands far more endurance than that. Smile and have a good time anyway.

Language models and coding agents are great, but there is more to life, and more to AI, than just LLM agents.

One thing I didn’t appreciate- humans are somehow very attuned to slight variances in people. Seeing the same shit everywhere make us feel crazy.

I’m pretty optimistic that AI agents will replace humans on tasks and not roles. LLMs are fundamentally constrained and do not have a world model that understands physics and deductive reasoning. Thus they exhibit jagged intelligence with hits and misses.

One of the worst parts of corporate life is the completely unnecessary urgency around everything.

Having the bravery to move from your birthplace/family is so rare.

Man we are so tired of AI replies lol

Just saving this here to document a story and as a self reflection on whether AI is really making me more productive Yesterday morning I found a way to complete the new HVM approach, that is much faster than before. I spent a few hours writing a spec, and then used Opus to implement. About 3k lines of C code later, everything worked and performance was incredible: 5x faster than HVM4 (stable at ~10x now). So, in one day I had outclassed HVM4. Incredible. I'd never have implemented that so fast manually. Now, enter today. I want to turn this into a real thing, but I haven't fully read the 3k lines yet. So, how do I trust it? I spent the whole day auditing the code. With AI. Several bugs found, most minor like forgetting to collect() some argument. But then I stumble upon this: λ{ inl: 1 ; inr: 1 } This was a test. But wait. This is matching on inl/inr. So the branches should receive the value of the Either. But they were numbers instead. Numbers aren't functions. This makes no sense. So why this is a test? It then stuck me. The AI completely misunderstood how function arities work. It literally assumed for no good reason that HVM5 was supposed to handle under/over-applied functions. For no good reason. I never wrote that. It never asked either. It just kinda thought "HVM is weird in some aspects, this might be one of them..." - and then it went on to implement a massive system to handle cases that should never happen to begin with. And all of that code is obviously wrong because it should not even exist. It is wrong. It is damage. And it is there. But it isn't too bad either. I just told Opus that it was wrong. Perhaps not so politely. And it solved it just fine. But then this begs the question. I spent ~20 hours in this file, and it is STILL not done. I went from 0 to 95% in the first 5 hours. Yet, 15 hours later, it is still not 100%. I suppose that is the real effect of using AI. If I had just written the C file manually in the last two days, would I not be further than where I am *right now*? Surely, the first version would have taken much longer to drop. But when I'd finish writing all that code, there would be zero, literally zero retarded shit. And, just today, I caught 5 or 6 retarded shit. And the worst part is: I don't know what the number of retarded shit left is, but I'm afraid it is >0. So if I have to read it all, review it all to ensure there is no retarded shit... what did I achieve by using AI, other than that dopamine anticipation?

i really like corgis. they did not deserve to become the mascot of grindmaxxing discourse.

Reservist on weekends is just weird. Makes the week feel so much longer.

@vikhyatk Sup

ai slop replies on this app have gotten really bad. i haven't read a human-generated reply in months

Been on Hermes Agent for a week- still trying to figure out the best use cases for it. I’ve been using it to prep for FDE roles lol.

@iScienceLuvr Any tips to validate if there’s alpha in AI x psychiatry?

a surprising amount of people don't understand the difference between AI in biology and AI in medicine... i've had to explain it countless times in the past ~4 years AI in bio - models for DNA/RNA/proteins/cells, AlphaFold, GPT-Rosalind, etc. AI in medicine - models for clinical data, patients, clinicians, MedGemma, ChatGPT for Health, etc. Lots of frontier labs are investing heavily in the former. At @SophontAI and @MedARC_AI, we work on the latter.

@hthieblot I’ve been finding the algo posts give better short term engagement but less takeaways. Switching back to my timeline

It’s my most insightful tweets that gets the least view. Algo sucks

What are some large corporations that have deep intellectual honesty as well?

Shots fired at Singtel 😂

Especially true in consulting- we use buzz words to sell to C suite, but forget that people on the ground who will actually use our solutions

Unpopular opinion: "AI makes everyone a developer" is true the same way "cameras makes everyone a photographer"