Kurzweil said AGI by 2029, Singularity by 2045.
It’s late 2026: models pass the bar, ship production code, and OpenAI says one solved one of the millennium prize problems. AI is superhuman in narrow spots, not ASI yet.
When do you think we reach ASI?
@JeffBohren The real concern is not passive LLMs (which lack goals or agency) but AI agents that use those models to plan, act, and interact with the world autonomously.
Partially accurate on market dynamics and incentives, but overstated and incomplete on motives, capabilities, and outcomes. The post captures a real competitive shift and legitimate skepticism about regulatory motives, yet it flattens genuine safety concerns, invents panic over IPOs, and exaggerates the performance/price gap.
I wish people would notice that the safety scare is not driven by people outside of the big labs, but directly from their leadership. The "whistleblowers" are plants that are amplified from within the labs, and supported by the CEOs. The goal is to capture the regulators, which will not shut down the big labs, but turn them into an oligopoly and outlaw decentralized AI.
@levelsio recalibrate your priors because if claude thinks something takes a day and finishes it in five minutes you just accidentally hired the most sandbagging engineer on your team
@BenjaminDEKR because knowing what you want and knowing what keeps you scrolling are two completely different optimization functions and only one of them makes money
@thedankoe this is good advice and the fastest way to feel even more lost is to spend six months researching the perfect thing to build instead of just building it
@neetcode1 it is not a skill issue it is a trust issue and the fix is the same as it always was which is you have to read every single line like you wrote it yourself even when you did not