Mizza

Less mathy discussion on Twitter, that also covers some of the speedup that comes from outside the optimiser https://x.com/classiclarryd/status/2104688738354028962

Massive result

k8sification of AI was always inevitable, if only as a form of salary justification.

It's as if two private companies are each building increasingly large nuclear bombs, both saying they'd love to stop but it would be unsafe to let any one company be in control of the nukes.

If you'd learn about a proposed experiment to test the quantum nature of gravity, this is a cool video:

https://www.youtube.com/watch?v=Uey_mUy1vN0

Hopefully we'll see a result in the next decade

The story of the High Line in NYC is pretty similar, Burnside in Portland and FDR in Philly.. you can just do things! (Except in the UK, where you can't.)

Libdragon has been improved massively in the past year or so, there are now simple demo games that show capabilities far beyond the original N64 games (bump mapping textures, HDR, seamless loading of massive overworld maps, etc.) Really talented core team, fun to give them a follow.

Demos: https://www.youtube.com/watch?v=XP8g2ngHftY

Equally as trashy and sick as Bumfights, the sort of thing that might happen in the early scenes of a documentary exposing the rise and fall of another fraudulent startup. I'm sure he'll be in the 30 under 30 soon.

Ouch, that hits close to home, and it seems like it does for a lot of others out there as well.

So what's the solution? Is there a playbook that avoids these pitfalls, or is it just the cost of the spin. Ideally, something early engineers can point to when we see non-technical founders falling into familiar traps.

Are you talking about Zig's MultiArrayList?

It's up to you to choose to who to believe about what happened on the yacht in the gulf of Mexico, the seventy year old billionaire cheating on his wife, or the thirty year old woman who accused him of forcible rape. There was some form of financial settlement, but it's still in private arbitration.

The public documents said: “He followed me into a shower, slammed me against the wall, and forcibly raped me. I begged him to stop and cried out that he was hurting me, but he ignored my pleas. The next morning, Schmidt attempted to convince me that I enjoyed the assault.”

The point is that student activists handed out flyers about this in advance, so the crowd was aware.

Eric Schmidt also raped a woman forty years younger than him, the students were objecting to that as much as the AI. Maybe don't schedule public speaking events after being accused of rape if you don't want to get booed.

If you found Phish twinned well with deep, focused work, you might enjoy grindcore for frequently interrupted work, since most of the songs are less than a minute long, like these: https://discordanceaxis.bandcamp.com/album/original-sound-ve...

I've got some projects that use ADCs on Pi "Hats" that connect to controls, I could see a future version which integrates that ADCs and pots directly onto the board to get a slimmer profile. It's quite handy, I wonder what the unit cost is with assembly at JLCPCB.

This is bonkers. Video on GitHub: https://www.youtube.com/watch?v=DDO2EvXyncE

I'm impressed by the kid's engineering and gumption, but I think he's a bit.. misguided, if you'll pardon the pun. The video ends with shots of Russian drone war, and, bizarrely, photos of David Koresh.

I don't think this ends well.

Reminds me of when every college kid had those plastic Macbooks. Quite a smart product.

Check out Seedance 2: https://seed.bytedance.com/en/seedance2_0

Nano Banana was technically impressive the first time, but after Seedance it's not really. It's all just an internet pollution machine anyway.

This is pretty wild! Only Llama3.1-8B, but this is only their first release so you can assume they're working on larger versions.

So what's the use case for an extremely fast small model? Structuring vast amounts of unstructured data, maybe? Put it in a little service droid so it doesn't need the cloud?

At this rate, I am expecting that an AI will be able to port the entire Linux kernel to Rust by the end of the year.

Would love to see them do another OSX 10.6 and just release a version with lots bug fixes and no new features. But instead it'll be a new half-baked LLM tool to help you make new half-baked LLM tools.

But that's for synths where the oscillators are ICs. I'm talking about simple oscillators from basic components

Never found it hard to build an oscillator, the hard part is musical voltage per octave. 3340 repro chips are the way to go, the best non-3340 circuit I've seen is this one and it's still temperature-sensitive: https://www.youtube.com/watch?v=FiCMjt0mqvI

Memory loss from sleep deprivation is an evolutionary advantage. If you remembered how rough the first few months of new children are, you wouldn't do it again.

This looks like dogshit.

I've seen a dozen frontpage posts this week that were nothing more than "look at this garbage Claude made for me." Can we get some new moderation rules to prevent slop on HN?

Reinventing the wheel and then charging for simple features?

If this paradigm excites you, just use Phoenix, dawg. It's so far ahead, everything else feels primitive.

Came here to say this. I use fastmail am quite happy with it because I just want a reliable inbox and nothing else. Just keep it running and don't touch anything else.

Has Medium stopped working on Firefox for anybody else? Once the page is finished loading, it stops responding to scroll events.

WashPo article on this was interesting: https://www.washingtonpost.com/health/2025/09/22/leucovorin-...

Seems like it works only for a very specific type of childhood autism, but if my child had this I would be kicking down doors to get it. The article has some good insight into how honest researchers feel about their work being trumpeted by the scientifically illiterate carnival barkers in charge of things.

Even at $10k/yr/employee, you'd need 30 million people on the 10k/yr plan to hit 300B ARR. I think that's a hell of a big swing. 3 million, recoup over ten years? Maybe, but I still don't think so. And then competition between 4 or 5 vendors, larger customers figuring out it's cheaper to train their own models for one thing that gives them 90% of the productivity gains, etc.

But rather than speculating, I'm generally curious what the companies are saying to their investors about the matter.

What's the path to recouping that money?

Even if every major company in the US spends $100,000 a year on subscriptions and every household spends $20/month, it still doesn't seem like enough return on investment when you factor in inference costs and all the other overhead.

New medical discoveries, maybe? I saw OpenAI's announcement about gpt-bio and iPSCs which was pretty amazing, but there's a very long gap between that and commercialization.

I'm just wondering what the plan is.

Until there's a paradigm shift and we get data and instructions in different bands, I don't see how it can get better over time.

It's like we've decided to build the foundation of the next ten years of technology in unescaped PHP. There are ways to make it work, but it's not the easiest path, and since the whole purpose of the AI initiative seems to be to promote developer laziness, I think there are bigger fuck-ups yet to come.