Every thread apart from this one is babbling on about something irrelevant. These next few months will be the most consequential in human history, and not because of Iran of because of crpytocurrencyLast November, AI was a barely useful search engine supplement. Today, only a few months later, it's more capable than new grads in white collar professions. The productivity gains/economic impact of this is inbound (we're in the electric motor attached to a line shaft phase of this)And yet it shows no sign of slowing down. Data center construction has just overtaken office construction for the first time in human history. These next few models will show if our outcome is only white collar layoffs/mass unemployment, or something much more strange.How are you pricing this in? I expect a collapse in luxury, travel, software, I expect large scale mortgate defaults. How about you anon?
>>61936581buy an ad
>>61936581>SAARYou guys arent even trying to hide it
my boss has been trying to automate my job for 2 years and he's getting no where. its kinda sad honestly.
>>61936587you are a nattering reflexive formulaic loser, but so am i, we are all about to be nattering and straightforwards
>>61936581what if the bubble *pops*?
>>61936602go back, begone
>>61936609I want you to be right. I love my job. But it turns out intelligence was never important for many highly paid jobs. We were all just taking text in and pushing text out the whole timeI hope you're right and that we're half way through a sigmoid
>>61936597Something has changed with the most recent frontier models. If your boss was trying to replace you with models from 2024 and before, they were insaneBut this is the nature of exponential growth. Nothing, and then bang, you're obsolete in the blink of an eye
Read this Stanford paper on a Computer Science problem Claude just solved and tell me you think AI is still stupid. I bet 99% of the people on this board can't even understand the material in the paper. https://cs.stanford.edu/~knuth/papers/claude-cycles.pdf>inb4 moved goalposting
>>61936677It's a bit much to call this a paper (its a loaded term in academia)But its worth nothing that Knuth is one of the most intelligent and capable computer scientists of all time, and also that he was an AI sceptic until recently, just like me (the last bit)
>>61936581>Today, only a few months later, it's more capable than new grads in white collar professions.is it now?>>61936677I'm gonna spell it out for you: kids at Stanford are also competing for internships
>>61936718What's your point about kids at Stanford
>>61936581I biteWhere can i invest?
>>61936788That's what im trying to figure out...If things are reaching the end/labs are brute forcing feigned progress by scaling compute (see gpt5.4s rumored "extreme thinking" (gay name), with anthropic rumoured to have something similar where they run a model in the cloud for weeks), then the economy will remain recognisable, although with lots of unemployment/greater concentration of wealth.But in the worst case, where human intelligence becomes obsolete? I have no idea, maybe you can tell me?
>>61936817If human intelligence becomes "obsolete" we have a literal existential crisis on our hands. We either submit (and die), or we destroy the new tower of babylon
>>61936650Which models would these be? I have a subscription to ChatGPT and it’s just barely worth it. Barely useful. It makes so many mistakes and doesn’t reason very well. Honestly don’t see it replacing many people.
>>61936914I started using Opus 4.6/whatever the new Sonnet is at work, in claude code. I work on a complicated code base (bespoke C++, not webdev)
kind of retarded question but does anyone know any good youtube channels or podcasts for keeping up with this shit, i feel like i’m always month or so behind. i had no idea claude was anything other than a joke until a week ago
>>61936914You should keep this information to yourself, just indicates you’re low IQ. The output is reflexive of your midwit input
>>61936817if unemployment reaches any kind of sustained double digits i expect UBI or something like it, which will cause inflation. so assets must be a good investment
>>61936939I wouldn't watch youtube stuff, id just keep an eye on these benchmarksmetr time horizons (going exponential)Gdpval (how useful they are at completing realistic tasks, >70% saturated)And i think most importantly, the remote work index, meant to measure how much an AI can substitute a white collar remote worker. Last measured model was around 3%, up from 0.8% a few months ago. For some reason no one talks about this one, because 3% doesnt fit the hype, but its the most important benchmark imo
>>61936984interesting resources, i will look at them. ty anon
>>61936947Your post seems randomly angry and attacking>>61936951If this is just the beginning, then there will be UBI/huge taxes, which sucks and is the best case
>>61936984It feels like every few months some benchmark gets quietly unfollowed as it doesnt grow anymore and some new shiny benchmark hype replaces it
>>61936947The smartest people I know say the same thing though. You have to put so much effort into goading the right answer out of it you may as well have done the work yourself.
>>61936914Chat got and copilot are dogshit, Claude is slightly better.
>>61937195>It feels like every few months some benchmark gets quietly unfollowed as it doesnt grow anymore and some new shiny benchmark hype replaces itthat's because the benchmarks are being saturated, you can't exactly grow much when you are hitting 90%+ on them
>>61937195rli is the only benchmark that matters
>>61936817Saar bnkr coin is building the rails for the agentic financial system. AI+crypto is a given. Also your mother tongues my balls
