ForHackernews

Possibly related to https://already-clean.goodfire.com/

Possibly. 1.58-bit models already run on CPUs https://github.com/microsoft/BitNet They're not good exactly but they are surprisingly better than you might think.

https://huggingface.co/microsoft/bitnet-b1.58-2B-4T#evaluati...

"It's not about how much you earn. It's about what you're worth. And who's worth the most? Companies that lose money." ~ Russ Hanneman

Meta are using AI internally to target and sell ads more effectively.

Consider: if you had built a hyperintelligent stock-trading AI, would you:

- Charge $0.70/M tokens for other people to ask the bot what stocks to buy.

- Get rich following the bot's trade advice and never tell anyone about your bot.

I use Claude Opus at work and DeepSeek v4.1 Flash on my personal projects. For the kind of work I'm doing, they're basically interchangeable and DeepSeek costs 3% what Claude does.

https://openrouter.ai/compare/anthropic/claude-opus-5.5/deep...

This turnaround time is appalling, especially given that Debian shipped a fix weeks ago. Poor show, Canonical.

That's another way of saying the same thing: even if AI is great, there's all kinds of reasons why it won't be a new industrial revolution.

Intelligence is very rarely the limiting reagent in progress.

Probably wildly irresponsible given they think it's likely to kill everyone with a virus, but yes.

If you have a super-intelligent AI that can cure cancer, just sell the cure for cancer already?

I'm not sure I'm any faster, but I think the end result is better and I've attempted some big refactors and new architectures that I would never have tried without AI.

N=1 but I'm not sure it's made my bosses any richer. What we're shipping has better tests and documentation than it ever would have before AI.

I'm not saying it will "fall apart", I just think the burden of proof is on AI boosters for their extraordinary claims.

They are saying this thing will soon obsolete human thought; meanwhile I use these tools every day and can't get it to refactor a React app cleanly without making a mess.

It's telling that OpenAI's decided they're going to sell ads in ChatGPT. Instead of, you know, using their AI themselves to produce all these valuable and revolutionary breakthroughs they claim it's capable of making.

We invested at least a couple trillion of speculative value into making ugly cartoon apes the future of art, finance and property. I'm not sure it's always possible to overwhelm reality with sufficient money.

Why do you believe that LLMs are good at the same things humans are?

Does this mean Caddy is vulnerable? IIRC, I think it's written in Go.

Builders who want a terribly inefficient and slow emulator for old silicon?

I assume that "defensive AI" means finding and patching vulnerabilities before attackers do, not some kind of live LLM packet-inspection.

It is possible to write more secure software[0], companies just choose not to prioritize security.

[0] https://cr.yp.to/qmail/qmailsec-20071101.pdf

What's the mitigation for this? There's no patches available yet https://ubuntu.com/security/CVE-2026-98432

Oh, that's good to hear. Can you move an existing account?

I'm also a happy Fastmail customer, but they are not really comparable: Fastmail is an Australian company with American servers, they are not a privacy-conscious option at all.

They should make a law against giving psychedelics to Bay Area rationalists.

Homer: Kids, there's three ways to do things: the right way, the wrong way and the Max Power way!

Bart: Isn't that the wrong way?

Homer: Yeah, but faster!

Speaking of cherry-picking, props for avoiding the entire point of my comment and instead pulling out some carefully-chosen NUMBER GO UP examples.

With apologies to Mr. Buffett[0] if you, @anonym29, could wave a magic wand and choose to own either...

A) All 21 million bitcoins or

B) All farmland in the United States.

I submit that only an idiot would pick A.

[0] https://ivanhoff.com/2013/04/15/warren-buffett-on-gold/

I feel like you're moving the goalposts slightly: Yes, all technology involves prototypes, development, gradual refinement to something production-worthy and offered for sale.

LaserDisc is an interesting case: is it failed proto-DVD, or its own separate thing? (analog, more like an LP than CDs or DVDs)

Not all technological innovations involves a failed attempt at commercialization. Microsoft really did try (and fail) to sell tablet devices with WindowsXP, https://en.wikipedia.org/wiki/Microsoft_Tablet_PC it wasn't just a prototype that they experimented with and then shelved until it could be refined later.

Why do cryptobros always refuse to admit investing exists? Forex and crypto trading are always zero sum, but you don't have to choose between dollars under your mattress or bitcoin on a thumb drive, you can actually invest in productive assets.

SPY has outperformed BTC over the last 5 years, and with far less volatility.

Bizarre take:

> The USA needs an equivalent of China's State Administration for Market Regulation (SAMR). This regulatory body generally prevents this tech platform monpoly nonsense

I guess the author is completely unfamiliar with https://en.wikipedia.org/wiki/United_States_antitrust_law

Firefox/Ubuntu I can't seem to move my farmer: clicking does nothing and arrow keys also don't move him.

There are non-software jobs, and also paid work supporting and developing free software.

With the arrival of AI, many of us will need to career change anyway.

I think the IRS already did some version of this and open sourced it https://github.com/IRS-Public/fact-graph

There's an article about it here https://biggo.com/news/202510160717_IRS-Fact-Graph-Public-Re...

These were artifacts from the IRS Direct File program (RIP)

Maybe your project could (re)productionize this dataset.