News for hackers. Stuff that matters.

tsunamifury

This is like how soldiers use tools worth orders of magnitude more expensive than their own pay.

It’s not weird once you think about it. Money doesn’t create an operator. But operators can be cheaper than tools and often are.

Re:Astra can read and write weights to improve Qwen on Monday November 02, 2026 @03:03PM (#50342842)
The fundamental flaw in the SOTA houses is that knowledge must be transitive or it isn’t useful. And if it’s transitive, it’s not moat worthy.
Re:We Took over ChatGPT Accounts on Saturday October 31, 2026 @11:09AM (#50316868)
I think you know that the mentality of a company like this is “security is a risk but not a reward and whoever captures the reward kills everyone else”
So you are saying an order was established...

I think you're the type of person that when wrong sorta talks about a bunch of adjacent topics to try to avoid the point of admiting that you were wrong.

The Sino-Soviet bloc and the Warsaw Pact controlled roughly a third of the world's population, an immense percentage of the Earth's landmass, and practically every natural resource required for autarky. They had nuclear parity, massive industrial bases, and their own closed global trading network (COMECON). If a system with that level of geographic, demographic, and military shielding cannot sustain itself, it is not due to a lack of opportunity.

It is prima-facie absurd to claim "it just wasn't given a chance"

I can understand your perspective here, and argue that, ultimately a market order was imposed on the world by America (and England) -- even when America interefered, removed, or otherwise altered democratic trajectories still resulted in a sort of networked world order. I'll not argue that they were not done, or even that they were done to benefit those targeted.

Simply that a type of order was enforced to build a global money-ist theory of society and and market + trade platform.

I might, wrongly, argue that many of these regions both would have continued in the chaos and death they otherwise were already participating in, but i'd also accept that the Dutch in Africa, English in the Caribbean and China, etc all introduced net new outside chaotic aspects on purpose. I'd also accept that attempts at order by the American's ultimately resulted in chaos, especially of course in the middle east.

As to south america, this I'll acknowledge of course was an history of interference, one that is hard to argue the contra-positive since communism in almost all regions of the world failed to develope a long term stable order which was both culturally, trade, and fiscally sustainable, with many falling to a worse state after these outcomes.

Re:OpenAI's agents talked to each other through the CPU Cache on Thursday October 29, 2026 @11:27AM (#50290612)
What?!

This is the most isane nonse I’ve ever heard. Heat cpu control has nowhere near the accuracy or speed to communicate anything meaningful without a predefined code, sensors system, and of course the accuracy noise of the signal would deeply be in question.

Who wrote this. This is prima-facie bullshit.

I don’t think you need to be so gentle with us.

We, as Americans have had our “mask off” moment now and are openly declaring ourselves the preeminent empire on the planet. It’s simply that we are withdrawing from our globalist support structures due to a wide variety of forecast reasons along side that, which is confusing to many.

It was our liberal values and performance internally by the Democrats that gave our people as sort of comforting story before. We do it for the good of others we give we topple dictators etc.

I am a deeply “real politik” person and I don’t think you average right leaning American ever thought otherwise. It was somewhat ironically the hope yet no-change left Americans (who i have counted myself among) who tried to give it the thinnest veneer of responsibility.

The greatest thing we provided the world was order, for a time. That seems true of almost any mature empire.

Re:Genie 4 on Tuesday October 27, 2026 @03:42PM (#50264683)
People do it in all forms of “cinematic world building” including amusement parks, vacations, games books movies etc.

Reality is very much blurry

Re:Genie 4 on Tuesday October 27, 2026 @03:41PM (#50264665)
Ultra nerdy point: in the matrix it’s revealed that humans did in fact choose to join it. And the meta point was that most of them still would choose it over reality [1] every day.

[1] reality is also somewhat in question in the films universe

Re:Anthropic S-1 on Monday October 26, 2026 @09:12PM (#50253511)
They could. Let’s see if they do. I’m just saying it ytd if they are yahoo or google.
Re:Anthropic S-1 on Monday October 26, 2026 @03:51PM (#50249934)
If you'd ever worked for Google, you'd know how insanely you're missing the point. There have been probably over 1000 self disrupting changes to google search as it continued to stay on top.
Re:Anthropic S-1 on Monday October 26, 2026 @02:24PM (#50248530)
either ... portability is nearly instant except for KV Cache, which is extrodinarily expensive to store.
Re:Anthropic S-1 on Monday October 26, 2026 @12:57PM (#50247027)
"Yahoo! is the best search engine, and I cant imagine anyone making anything better."
Re:Anthropic S-1 on Monday October 26, 2026 @12:56PM (#50246999)
So far anthropic can't 1) grow food 2) build things 3) do anything physical. Software engineers often fatally misunderstand that their code does nothing, it just coordinates.

You might lose your value, others might gain.

Re:Anthropic S-1 on Monday October 26, 2026 @12:48PM (#50246876)
No, they specifically are not sticky at all. Thats the whole problem.
Re:Anthropic S-1 on Monday October 26, 2026 @12:47PM (#50246857)
This IPO, if successful, will complete rewrite the standing rules of the stock market in ways incredibly harmful to the future of American style capitalism:

1. The Separation of Economic Value and Corporate Governance creating The Multi-Class Autocracy

2. Institutionalizing the "Off-Market Vault" that can manipulate stock price at any time

3. The "Vendor-Financed" Public Offering (Circular Balance-Sheet Refinancing)

4. Monetizing Non-Cash Derivative Liabilities as "Losses" Without Cash Burn

Re:Today, we hacked what we'd set out to hack: the iPhone on Saturday October 24, 2026 @11:49AM (#50222433)
Plenty of small companies do 1-4. Its customer oriented for profit vs hobbyists and zealots
Re:I Don't Understand Quantum Physics (2018) on Saturday October 24, 2026 @11:15AM (#50222087)
It’s easier the accept quantum physics if you first

1) Disregard many worlds as a purely remaining mathematical outcome of bounded probability rather than discrete reality. (No 1 thing exists just 1ish things but ultimately that doesn’t mean there are many versions of any give 1)

2) just accept you can’t find the “stuff” of reality. You simply measure increasingly small probability domains which interact with other probability domains

3) time is the sequential working out of probability of anything existing and other things that might exist co defining themselves.

4) space is just the interaction of these probabilities working themselves out into a stable projection.

Re:I Don't Understand Quantum Physics (2018) on Saturday October 24, 2026 @11:11AM (#50222042)
Many worlds is a mathematical construct of our inability to measure specific dimensions hyper locally and an artifact of that. Which for some reason the entire scientific community mistakes as a real observation of reality.

Just because at a quantum scale everything becomes potential doesn’t mean all potential has been manifested.

It baffles me that so many serious scientists don’t understand this basic tenet.

Re:Today, we hacked what we'd set out to hack: the iPhone on Saturday October 24, 2026 @11:05AM (#50221976)
Your opinion doesn’t really matter macro. The market has spoken.

You’re literally proving my point. This is built for yourselves not others.

Re:Today, we hacked what we'd set out to hack: the iPhone on Friday October 23, 2026 @01:17PM (#50212356)
We gave in because open source and hobbiests do not design good products and even worse shared systems, as much as they enjoy self-righteously declaring their values. And shared systems require trust, planning, and goal setting in a way no open system would ever build.

1) They of course lack the resources

2) Often lack professional displince across all skills required

3) Mostly make for themslves or whats interesting, not whats needed.

4) The release schedule is laughable at best.

Why doesn’t your lawyer just say “asked and answered”
Re:Publish fake stats, get cited in LLMs on Tuesday October 20, 2026 @06:28PM (#50173568)
You’re talking to the guy that invented next word and phrase token prediction at Google.

Gullible is an appropriate colloquialism here. As in they don’t check countervailing narratives naturally and fall into a recency sink on last context. Long walk context construction is hard, time consuming and most Importantly significantly more expensive.

Re:Publish fake stats, get cited in LLMs on Tuesday October 20, 2026 @12:52PM (#50169240)
Yea if you ever want to test how gullible even latest SOTA models are ask it for financial advice on a specific IPO and watch it vaccume up sell side nonsense and spew it out back to you as absolute fact.
Re:Publish fake stats, get cited in LLMs on Tuesday October 20, 2026 @12:51PM (#50169222)
it seems like end stage of LLMs is an MCP economy in which Google and Apple will have a massive advantage in verified locations, businesses, identities and representation of services accurately.
Re:Facebook now charging to post more than 2 links a month on Tuesday October 20, 2026 @12:04PM (#50168433)
This is a bi-directional play. For the vast majority that dont, this is an attempt to get more content INSIDE the facebook ecosystem as a moat, and for those that need outside content links they will pay since its a organic premium conversion.

Not defending it, just explaining the logic and why it might work even if it seems superficially stupid.

Re:DeepSeek just open-sourced its inference engine on Tuesday October 20, 2026 @11:26AM (#50167845)
According to who? There is no agreed upon anyone that says that?
Re:DeepSeek just open-sourced its inference engine on Tuesday October 20, 2026 @11:22AM (#50167786)
I think it’s pretty clear from the cherry picked numbers at Anthropic.

Demand is the issue. Demand only spikes with a sota model release then in drops back down to unprofitable levels when factoring in training a new model and build out.

So your stuck in an unprofitable cycle of losing revenue when your not releasing sotas but losing profit when you are.

This goes back to googles original memo “there is no moat”.

Where in you can’t stop training sotas because you’ll lose customers to the next inference model above you instantly.

Google has seemingly defeated this by finding models cheap and good enough for their existing platforms resulting in a moat of distribution. Something OpenAI and Anthropic seems to lack.

Re:Making Games with my 4 year old son and Claude on Saturday October 17, 2026 @12:06PM (#50132294)
Yea we do that as well. Why assume we wouldn’t do something in real life. This was an illustration.
Re:Making Games with my 4 year old son and Claude on Friday October 16, 2026 @02:58PM (#50122839)
Last weekend my son asked me how trains worked, and I decided it would be fun for us to find out together. We sat down and I made a small train simulation showing him track switching in Switzerland. He enjoyed it so much we kept building and adding new ideas and assets until we ended up with Tiny Worlds, miniture set of games set around transit.

Anyways 10 hours later Dad has a new hobby and my son has moved on. So I'm sharing it here for fun and feedback.

Older »