OpenAI and Anthropic are not Amazon. They are AOL
LLMs are already close to being fully commoditized
The giants are yet to come. They will be made by startups working at a layer below the models themselves
The frontier LLM labs are doomed
Did anyone stop to think that maybe they’re just heinously overcharging on API rates?
Compare it to deepseek et al to get a comparison for how much inference actually costs them
Replying to @SemiAnalysis_
Recently, we purchased one of each Anthropic/OpenAI subscription plan and randomly ran long horizon coding tasks until we exhausted the weekly limit. It's widely believed that a $200/month plan maxes out at ~$2000/month worth of tokens (assuming API pricing). However, we found that the subscriptions are actually far more generous. (2/4)
Claude Fable, like most LLMs, is made to waste tokens. It justifies more business investment and expansion
To get around 10x more usage, just ask it to delegate work to agents that have the minimal context they need
It saves dramatically on input tokens
Benedict retweeted
Install Disk Creator 2.0 facilite encore plus la création de clé d'installation de macOS macg.co/logiciels/2025/09/in…
New Install Disk Creator update just released
Finally notarized it now because XCode wouldn’t let me compile it for the last few years
What if we get artificial super intelligence that develops sentience and it does go rogue and kills people. But only the really bad people that are screwing the world up, because it’s really cool and smart, and it acts nicely towards everyone else?
I’m blocked by him but…
@IfindRetards
This is incredibly racist. Those poor women. Who put them in that capsule and launched into space? NO NEED TO DO THIS. This could have just been tested on robots
The current wave of weird events in politics only make sense when you look at everything as a race to build the first robot army
We’re in the Manhattan project 2.0
Everyone talking about misaligned AI but nobody talking about the next game theory rational move
Be the first team to make a robot army and kill everyone else so they don’t get their move
Elon Musk featured by Jerry Springer
Donald Trump featured by The Apprentice
Political theatre brought to you by the entertainment branch of the military industrial complex
The common Whisper text to speech model and its embeddings appears to directly mirror how our neurons handle it. This is nuts
Inspired by the success of LLMs, today on the blog we discuss how neural activity in the human brain aligns linearly with the internal contextual embeddings of speech and language within LLMs as they process everyday conversations. Learn more →goo.gle/4iiUoNj
There’s two kinds of people. There are above-average people who are busy doing things like building stuff, building companies, doing research, doing science, etc. And below-average people who are busy sitting around forming committees to manage and manipulate the above-average people. So they end up falling into positions like being in the government and other kind of managerial positions where talent and competence are not required. And then what ends up happening over time is that below-average people actually tend to gain the power since they're focused on just grifting and finding ways to manipulate the system while all the above-average people are busy actually doing stuff. And that ends up becoming the basis for governments
The idea of universal basic income cracks me up
They’re already getting to kill us while we are productive members of society
Can you imagine how much they’re going to kill you if you’re unproductive???















