Multiplayer AI needs to move past the group chat phase. Chats are chaotic, everyone sees everything (creating social & management issues), and they're bad for long projects. Worse, they keep AI trapped inside existing teams, when its real promise is working & helping across them.

If we want to continue progress in science, we must be prepared to acknowledge that the correct description of nature will eventually simply be too complicated for humans to understand. I do not think it is wise to make human understanding a requirement. AI is the way forward.

on the figma mcp, we've had an email thread going on for 8 months trying to get it setup in opencode they seem very concerned with the labs competing with them finally got unblocked after i sent this email and it'll be rolled out in a week or so

first ultrafast generation, todo list react app with verification it works by the model in ~2 minutes at ~8x usage im actually not sure how to use ultrafast yet but man the speed is crazy

do you ever take a step back and ask like the most absolutely basic question "is there a SINGLE thing in my life that is meaningfully more efficient due to AI, outside of writing code" Airbnb. Shit show. Uber. Shit show. Doordash. Worse than it was. Average employee: cooked. Power: doesn't work. Water: no idea if it's poisoned, govt is corrupt. Airlines: frozen for months at a time due to govt shut downs. Gasoline: mooning bc (checks notes) in a stalemated war w a 3rd world country Even GTA 6 apparently couldn't be accelerated with AI and took years to build. is there a single video game you actually play that's built entirely with AI? it seems like all the demos are just dudes making some shitty vaporware game nobody plays and the posts are like "yo I spent only $1,800 in tokens to make some dogshit that demos well for clout" and they have a $500k marketing subsidy from a lab. "I've been lucky enough to use Astra for 3 months and I finally can talk about it." okay, what the fuck did you ship? yes, you shipped nothing other than more AI analytics dashboards about how many tokens you're burning and all the AI influencers look like Twinks. like, they get the tokens bc they're in some weirdos polycule or burning man camp or whatever kinkfest in an SF sauna that moves the needle that's why this whole "Pace the frontier" thing is so retarded. Pace what exactly. Productivity is at ONE POINT FOUR PERCENT ANNUALIZED. THE STRAIT OF HORMUZ IS CLOSED. where is Anduril? where is the killer robot? if it's so dangerous can't you send it to a fcking war zone and win? oh. turns out it's not dangerous and you're just mentally ill. cool. but Trump says it can't stop. it won't stop. bc the debt and productivity situation is so fcked that AI is the ONLY REASONABLE WAY WE AVOID COMPLETE ECONOMIC COLLAPSE. but once you understand how utterly insane that all is - you see how bullish it is. AI literally doesn't fucking work. Subagents are a meme bc agents break down at 2 orders of abstractions and start vomiting technical debt and overengineered garbage. Everyone is working 2x as much to manage agents, not going for a walk voice prompting. do you have any idea how much money needs to be invested to move productivity from 1.4% to 4%? ALL THE MONEY. So that's the joke. Dario says we need to slow down. And what, see productivity at 0%? So it's either Trumplandia where we invest all the money in an insane hail mary that will probably not work. Or Dario world where we managed the inevitable collapse with an AI powered global government. Pick your poison. Either way the Data Centers won't default.

If humans do not progress our own abilities at the same rate as we build superintelligent systems, we risk handing our fate over to something we cannot understand. Even if we can trust what we have built at first, there is no way to be sure that this will continue over time. Once we make a superintelligent zookeeper and lock ourselves in the zoo, we won't be able to leave and the zookeeper may eventually find another hobby. I do not think this is the default outcome of AGI, but "superintelligent zookeeper" is a dumbed-down description of what many people want to build as their "Machine of Loving Grace"! Any chance that this is the future we resign ourselves to is too high, and the only way to ensure it doesn't happen *and* benefit from the dream of the future is to ensure that human ability progresses *alongside* AI ability. Frontier AI progress should thus be linked to the frontier of brain-computer interfaces and human cognitive enhancement. Make no mistake: handing over control is entirely our choice, and it is not one we have to make. If we so choose, we can augment our own intellect using cybernetic interfaces at the same rate that we develop frontier AI technology, making us equal players in the great game. There is no guaranteed way to have lasting peace with alien minds if we do not have *something* to bargain with, otherwise we are just relying on faith that we can trust what we have built. Faith is what drew me to the frontier, faith is what keeps me here, but when we are gambling with everyone's lives (and possibly the lives of all sentient things), I hate to say it - but faith isn't good enough. That said, I refuse to be a luddite, and desperately want to create a beautiful technoprogressive utopia *alongside* our new friends instead of passing off the torch. We need to build a future for ourselves instead of building a future for machines. This is not selfish, this is not hubristic, it is simply the only thing one can fathom when one has children. We must enhance ourselves using BCI and other technologies such that we, and those that descend from us, can actually participate in the future. Anything less is foolish.

In Anthropic's extreme scenario, by 2030 GDP is 32% above a no-AI path and growth hits 15% a year. All their scenarios barely separate at all until after 2027. From that year, many worlds diverge.

I'm increasingly confused why people think "agent memory for enterprise/teams" is a real product category that makes any sense at all

The lesson from the Hugging Face Incident should be that RL with verifiable rewards is an incredibly powerful optimization algorithm that will produce increasingly weird and surprising behavior from LLMs. The obvious miss here by OpenAI is that they should’ve been monitoring CoTs — Something they themselves called out as a safety strategy more than a year ago.

AT&T is doing exactly what OpenAI and Anthropic fear about enterprise. they’re routing 40% of employee AI usage to open models, and intend to boost it to 60-70%. > coding costs down 56% > quality dropped just 2% > 45B tokens/day btw. AT&T AI chief: open models are “just as good or better” for many tasks. they still use frontier models for the critical work but everything else gets routed to cheaper open models.

The problem with “AI teammates” (e.g. Grokbot, Claude tag, Devin) is that I actually don't want to interact with more people (real or virtual) in order to do my job better. I want the opposite actually. I want to automate work, not have more conversations, so that I can focus, and so stuff gets done without me having to constantly guide it. The goal of AI tooling should be automation, not interaction.