turned an old Apple Watch into an iPod yesterday big fan, building new iPod apps now

What Is To Be Done? I propose a simple plan: Big AI companies should be allowed to build AI as fast and aggressively as they can – but not allowed to achieve regulatory capture, not allowed to establish a government-protect cartel that is insulated from market competition due to incorrect claims of AI risk. This will maximize the technological and societal payoff from the amazing capabilities of these companies, which are jewels of modern capitalism. Startup AI companies should be allowed to build AI as fast and aggressively as they can. They should neither confront government-granted protection of big companies, nor should they receive government assistance. They should simply be allowed to compete. If and as startups don’t succeed, their presence in the market will also continuously motivate big companies to be their best – our economies and societies win either way. Open source AI should be allowed to freely proliferate and compete with both big AI companies and startups. There should be no regulatory barriers to open source whatsoever. Even when open source does not beat companies, its widespread availability is a boon to students all over the world who want to learn how to build and use AI to become part of the technological future, and will ensure that AI is available to everyone who can benefit from it no matter who they are or how much money they have. To offset the risk of bad people doing bad things with AI, governments working in partnership with the private sector should vigorously engage in each area of potential risk to use AI to maximize society’s defensive capabilities. This shouldn’t be limited to AI-enabled risks but also more general problems such as malnutrition, disease, and climate. AI can be an incredibly powerful tool for solving problems, and we should embrace it as such. To prevent the risk of China achieving global AI dominance, we should use the full power of our private sector, our scientific establishment, and our governments in concert to drive American and Western AI to absolute global dominance, including ultimately inside China itself. We win, they lose. And that is how we use AI to save the world. It’s time to build.

people think software's big advantage over hardware is fast iteration it is. but that's not the real gap in hardware half your parts are backordered for months. you can't pay your way out. no expedited shipping fixes a part that doesn't exist yet. you just wait biggest opportunity in hardware right now might just be adding supply

Feel like the only real contenders to build 24/7 agents for mass market consumers are Google and Apple Google already has all the data and users trust Apple and will connect new sources to it

Happen to be intimately familiar with how hard PCB routing is, was unsolved on all models so far This is big if true Will go unnoticed for a bit and then suddenly be huge

I remember when chatgpt 2 came out in 2019 it broke the internet. just the output of natural sounding text was mind blowing. it didn't write code or anything else. i remember talking about how cool it was at a house party in sf 7 years later, the same "natural' sounding text output is slop and extremely annoying to read. i feel perceptible irritation when reading text that contains clear AI tells. demand for increasing levels of intelligence is infinite

> "agents are replacing engineers" not really yes, don't need an engineer to start something now; getting off the ground is available to anyone but keeping a large, living codebase healthy still takes real engineers; agents can't hold it together yet hard part has moved from writing code to maintaining it

update on @farcaster_xyz, @clanker_world and @neynarxyz we’ve started a process to look for a new home / team to run the products going forward. earlier this year, this seemed to be a good fit for our team: it was a developer-first network, and we'd been serving developers from the start. There’s been a lot of change since then and we’re not the right fit for what’s needed next. talking to a few teams we think could be a good fit for running decentralized social apps and related developer products. If you’ve thoughts / are interested, reach out. our company will return its balance sheet (most of it remains) and team members will move on to something new. this is not a decision we made lightly. I’m optimistic that with this next iteration, Farcaster will find its long term fit. it’s loved enough that it’s unlikely to simply perish - thanks to all of you, and lindy-ness.

gave my agents a standup channel so they'd coordinate, now they hold a daily standup this morning one apologized for being "away all weekend" ???

sama was right that AI is more like electricity than transistors there is a real continuous flow from a service provider to a consumer - model companies are the new utility companies pretty much all modern work stops when the flow stops, similar to the impact of power outages in the future, any model company outage will bring much of the physical world to a standstill calling it now

The Matrix except you can’t eject into a human (yet)

the year is 2032 AI does all the work we spend eight hours a day attending meetings about what AI did

one way ai has changed decision making at Neynar is that we can decide from prototypes instead of theory. we can turn a website redesign into a website 10 minutes into a discussion. the team can turn a feature doc into something usable the same day and ship it. sometimes it works out. sometimes the prototype can kill the idea. it's funny how there's almost no scope for disagreement on semantics and opinions once everyone can SEE something.

when bitcoin is up ~3% over the past month but you remember you bought at $82k

this is somewhat of a bad analogy but cheap intelligence might do to companies what consumer cameras + youtube did to media. before youtube and twitch, if you wanted to build a media career, you generally had to go through NBC or some other commercial aggregator. they gave you a slot or they didn't. then cameras became mainstream and distribution opened up. people could create a different kind of career and make money in ways that didn't exist before. NBC is still around as the corporate bigshot. but there are also far more small publishers than there used to be. I think something similar happens with companies. three years ago, if you wanted to start a company with a technical component, you were kind of out unless you knew how to build it or knew someone who did. today you can start. I don't think this would do any long term damage to established companies. it more likely means that there will be more small companies, more people trying things and a lot more products. most of the conversation today is very bleak about work AI will take away from people. but it'll also give people new things to do!

CLANKER JUST LAUNCHED A NEW LAUNCHER! and it's not just a token launcher it's a token launcher launcher lol launch your own token launcher in seconds or use someone else's to launch a token every token is paired with hoodie-2:native (!!!) yodawg.xyz by @clanker_world

recently, when we run into a problem at Neynar, I describe our current team + problem to the latest model and ask what we should do. it will generally say something like: hire someone for xyz. "okay, write the job description." it writes one. "now make an agent for the job description instead of hiring the person." you tricked it. then we start with the agent and see where it gets stuck. sometimes that point never comes because the agent is good enough and we keep using it. other times there are enough corner cases that it becomes clear someone should own it completely. even then, we now understand the role much better than we did when the model first told us to hire. QA has worked this way for us. agents scroll through our apps, tap buttons and report what breaks. it is not foolproof and a good QA engineer would still do better but we have not ended up hiring anyone for it yet. hiring is hard enough that starting with the agent has become a reasonable default.