Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Astra for Law(openai.com ↗)
    157comments
  2. Bend – A language that blocks AI mistakes via proof and runs on GPUs(bend-lang.com ↗)
    68comments
  3. I Hate You Microsoft(henriquenunez.eu ↗)
    19comments
  4. Bonsai 2 27B: Near-Lossless Compression in a 9x Smaller Footprint(prismml.com ↗)
    10comments
  5. Hister: A private search engine for the pages you visit and the files you keep(github.com/asciimoo ↗)
    115comments
  6. Wax motor(wikipedia.org ↗)
    30comments
  7. Fujitsu launches made-in-Japan next-generation CPU FUJITSU-MONAKA(global.fujitsu ↗)
    170comments
  8. Everybody's Lost Their Minds(netmeister.org ↗)
    123comments
  9. CrowdSec Source Code Leak(crowdsec.net ↗)
    33comments
  10. Rate limits on GitLab.com are changing(about.gitlab.com ↗)
    101comments
  11. Infinite-Parameter LLMs: Generating and Adapting Weights from Live Data(arxiv.org ↗)
    25comments
  12. Flet 1.0 – Build cross-platform apps in Python(flet.dev ↗)
    2comments
  13. Towards Self-Driving Codebases(detail.dev ↗)
    70comments
  14. TSMC revealing details about next gen A14 node(mapyourshow.com ↗)
    24comments
  15. The American Religion of Self-Storage Facilities(newyorker.com ↗)
    277comments
  16. Zettascale (YC S24) Is Hiring ASIC/FPGA Engineers to Build Chips for ASI(zscc.ai ↗)
    discuss
  17. Why I didn’t sign the Fields medallists’ letter(gowers.wordpress.com ↗)
    231comments
  18. How GLM built its own inference infrastructure(z.ai ↗)
    253comments
  19. How Uber Protects Against Retry Storms(uber.com ↗)
    discuss
  20. André Weil and the Hodge Conjecture(jiahao116.github.io ↗)
    7comments
  21. Running Ubuntu on the Lenovo IdeaPad Duet(vhaudiquet.fr ↗)
    15comments
  22. How do we prevent mathemathics from devolving into the Medieval Era of secrecy?(mathoverflow.net ↗)
    25comments
  23. GraphViz Pocket Reference – Make a Graph(grevian.org ↗)
    discuss
  24. One year of sponsored Servo development(servo.org ↗)
    136comments
  25. Launch HN: Skillsync (YC W26) – AI chat sessions made portable across agents
    42comments
  26. CCC invites all model citizens to 40C3(ccc.de ↗)
    171comments
  27. Canto: A speech model built for the real world(wisprflow.ai ↗)
    9comments
  28. Show HN: Share your AI Setup, Learn from others(mysetup.ai ↗)
    79comments
  29. The most important product decision is what you don't build(liamnugent.me ↗)
    3comments
  30. The Return of Sail Power: Cargo Ships Are Turning Back to the Wind(gcaptain.com ↗)
    119comments

Economic policy for AGI

56 pointsby 4h agoinstitute.deepmind.com
61 comments
3h agoHN ↗

My read, but the effective policies are infeasible, and the feasible policies are largely business as usual. Pretty much precisely as expected, honestly. I struggle to see how anything like this is going to be implemented without business as usual changing in ways that the current system is extremely resistant to.

3h agoHN ↗

I thought this was a good read with good data analysis, but at the end of the day none of this matters if the political and democratic systems collapse. If there is broad displacement by AI, I find it really difficult to believe folks will sit down and have a rational discussion about pros and cons of various policies, and it's nearly impossible to predict what will happen. I for one am fairly surprised that what I think is the valid anger among the US working class was successfully channeled into a cult around a grifter and huckster whose vast majority of policies are most detrimental to his biggest supporters.

3h agoHN ↗

I for one am fairly surprised that what I think is the valid anger among the US working

While surprising it should be too surprising. 1930-1945 was another example of this happening.

3h agoHN ↗

Though AGI isn’t here yet (and in my opinion will only be marginally useful until we have physical AGI), we need to speedrun the painful bits so we can get to the other side in one generation. We need to bear the brunt of it and be the sacrificial generation.

3h agoHN ↗

I think that there are a lot of people who would be unwilling to sacrifice their life.

3h agoHN ↗

Looking back at the beginning of the 1900s and now there are a lot of meta level parallels. Too much change is happening too quickly for societies to realistically adapt. This leaves a lot of room for greedy agents to gather other power seeking agents to further destabilize the situation for their own gain. The outcome of things like this is almost always war. Problem is once war starts it tends to spiral out of control and the body counts increase rapidly.

2h agoHN ↗

I would also add that the internet is very young and the amount of data and news we get daily is also new and different stressful than 50 years ago.

2h agoHN ↗

A real war would disrupt the internet so substantially as to render AI inoperative. Data centers are easy to find on Google Earth, few are really hardened against deliberate attack, and they would be immediate targets.

3h agoHN ↗

We need to bear the brunt of it and be the sacrificial generation.

Who is "We" here?

I'll take a guess: Gen Z, and Gen A.

Gen Z because they went through covid during college, and now this AI boom that's completely wreaked the entry level job market.

Gen A is looking increasingly grim, unfortunately. If any of you know any school teachers, you'd know how disheartening it is to try and teach someone that has a homework and entertainment machine in their pocket at all times.

Millenials went through everything Gen Z has and more ('08), but there was at least a few years of relative normalcy growing up. I don't know what's next, but I'm not very excited about it.

/rant

3h agoHN ↗

As always, named "generations" are completely arbitrary and meaningless. They're made up by marketers to sell products.

3h agoHN ↗

A generation is a time that people were born between and useful because it embodies much of the cultural context that people share.

3h agoHN ↗

No, it doesn't. The baby boomers represented an actual demographic change, that's why they're tracked by the Census Bureau. Everything since then is marketing and astrology.

2h agoHN ↗

It's the same, you're just substituting years for months.

2h agoHN ↗

The name of the generation gives you birth years. It also tells you "hey the generation 2010-2020 are getting hit by AI a lot more than the previous ones.

This not astrology at all. Astrology is 'magic thinking'.

2h agoHN ↗

Named generations are a marketing tool, it’s a concept as garbage as astrology

2h agoHN ↗

Its a name for a generation a decade wide. I really don't get your hate for it?!

2h agoHN ↗

BS. While I agree the cutoffs are arbitrary, marking a cohort by, say, "people who graduated highschool before the advent of smart phones" or "people who went through childhood in the early years of social media" are valid groupings.

As a Gen Xer my childhood was soooo different from what kids today experience. And literally every single Gen Xer I know is glad and feels relief (sometimes mixed in with remorse) that our childhood was before the consumer Internet took off.

2h agoHN ↗

If you want these cohorts to have statistical significance you have to add more details, such as the geographic location, language, etc. But you can create cohorts with pretty much anything, in itself that doesn’t validate the concept of generations as pitched by marketing agencies

1h agoHN ↗

Luckily for us, the cohorts don't need to have statistical significance because people aren't talking about statistics. They are talking about the broad shared experiences of most people born in a given range of years.

3h agoHN ↗

Correct. It is another way to divide people again, just like people who are pro-AI vs anti-AI.

Do not fall for it.

1h agoHN ↗

You are conflating "divide people" in the sense of "turn people against each other in order to divide and conquer" with "divide people" in the sense of "define groupings so that it is possible to have conversations about differing experiences". One is discussing a goal while the other is discussing a tool. Notably that tool does not necessarily need to be used for evil and often isn't.

1h agoHN ↗

Just because something is arbitrary doesn't mean it is meaningless or useless. In this context named generations are useful because they allow us to discuss the experiences of people born at different times. Certainly there are alternatives like going back to dividing things up by decades ("80s kid", "child of the 90s", etc.); however, that doesn't make named generations useless.

1h agoHN ↗

No it doesn't. I have little in common with a Japanese rice farmer born in the same year.

1h agoHN ↗

The named generations are generally used in the context of specifically American culture or at least English speaking culture. They were never intended to say you have generational similarity to someone in a completely different culture.

57m agoHN ↗

So now it only refers to Americans? The arbitrary goal post shifting continues.

37m agoHN ↗

It's not goal shifting, that's what it has always meant. I'm sorry that you didn't realize that.

2h agoHN ↗

Just another guy who loves humanity, but hates humans.

3h agoHN ↗

There's no reason we need to and there's every reason to believe that the people and companies foisting this technology on society will accrue the benefits and dump the fallout on everyone else.

2h agoHN ↗

There is absolutely no reason to have that being forced at full speed on the entire world without consent, no. That’s a completely insane thing to say

34m agoHN ↗

I don't think a sacrificial generation is necessary. There should be ways to sort things that are positive over all.

3h agoHN ↗

I'd rather read an analysis from the World's Socialist Website than from the perpetrators.

What are your concrete plans now for the 30,000 Oracle employees that have been fired because Larry got overly ambitious with AI?

2h agoHN ↗

You should want to read both sides and generate your own view of it.

You can't ignore signals of the frontier labs and people working there. How much you trust them and how serious you take them and how much it reflects on what you experience, is your personal analysis.

I'm playing around with fable 5.1 and Astra and its really crrazy how much better they got again now in comparison to a few month, i do trust them to a certain degree, that they are already seeing again progress internally we don't have yet.

The last update killed for sure a lot of low level/avg blender/3d artists.

2h agoHN ↗

Funny how Hyperscalers will write endlessly about this outcome but have no plan for what happens if LLMs turn out to be a bubble and the economy they're propping up goes into recession

2h agoHN ↗

What were your concrete plans then for the thousands of Oracle employees who were fired in 2005, 2006, 2009, 2017, and 2022 before they got ambitious with AI? Mass layoffs at Oracle are nothing new, they've never hesitated to make huge cuts.

Considering that socialism has been an abject failure every time it's been tried at scale, I fail to see any analysis from the World's Socialist Website could be useful.

1h agoHN ↗

Socialism has been hugely successful. Just look at Western Europe and the Nordic countries. Socialism has only been an abject failure if you swallow the conservative talking point that "socialism" is always equivalent to Stalinism/Maoism.

1h agoHN ↗

The Nordic countries abandoned their failed attempt at socialism decades ago.

1h agoHN ↗

I'm sorry but that's simply not true. They still have socialized medicine, generous social safety nets, significant worker protections, etc.

38m agoHN ↗

I think mixed economy socialism where some bits are socialised and others are not may be the way forward with AI.

1h agoHN ↗

There is some recent stuff on that website https://www.wsws.org/en/articles/2026/09/17/wrke-s17.html

Sanders and the trade union bureaucracy join hands with the fascist right

The question of who controls AI can be resolved only through the establishment of social ownership and democratic control over the technology by the working class, internationally.

the article also doesn't have much concrete plans for the Oracle workers.

3h agoHN ↗

Embodied robots is the most significant threat imo. Capital can be taxed but for now we have few general policy frameworks for the "who gets to walk and drive on the street" question

2h agoHN ↗

Scares me. When the ruling class get robots servants they will have no need for even simple human interaction anymore and that will truly unleash them to do harm.

3h agoHN ↗

Yet in the short term, significant transformations have often also produced lower wages for many, hollowed out regional economies, and deepened inequality.

"Short term" being rather loosely used? Most industrial regions have never recovered. Look at Gary Indiana, Detroit and surrounding towns, Toledo, other regions that were dependant on manufacturing or mining. Most have been stagnating for decades; few have found a path back to a thriving economy.

2h agoHN ↗

It's a multifactorial problem. The decline of heavy manufacturing was one cause of economic stagnation in Detroit. But the failed progressive social policies and lack of effective law enforcement didn't help either. Businesses are hesitant to expand in areas where they can't find reliable employees and where their property isn't secure.

2h agoHN ↗

Those places aren’t isolated though. Many of their residents (and especially their children) have left and found opportunities in other parts of the country.

Economists tend to prefer policies that help people, not places. Gary, Indiana just doesn’t have any geographical advantages that make it competitive.

I think, if anything, the mistake was to allow people in those places to believe there would be an eventual turnaround by trying to prop up uncompetitive manufacturing industries.

2h agoHN ↗

Gary, Indiana.

Large parts of Northern England.

Large parts of Scotland.

Walloonia in Belgium.

Northern France.

There are LOTS of regions where the people living there have never recovered after deindustrialization. We cannot expect millions of people to just uproot, it's not realistic.

2h agoHN ↗

This is true. Pretty much everyone I was friends with at school in the north of england left, most of us now in or around London. But we were the ones who went to university. Most who didn't stayed put.

2h agoHN ↗

Gary has a port on Lake Michigan, is right next to Chicago and its rail and transit hubs, is well-connected to the interstate highway system, has an airport... it's hard for me to understand how it doesn't have geographical advantages.

2h agoHN ↗

Gary has a huge geographical advantage in being located on the Great Lakes. This really cuts the cost of transport for bulk materials. And it gives them a reliable source of fresh water. It could be a great site for a nuclear power plant. Regardless of how AI economic policy shakes out, we're going to need a lot more electrical power.

1h agoHN ↗

What about the people who cannot leave?

17m agoHN ↗

Not saying these places will universally recover, or even any of them will. There are plenty of secular trends pushing industry to other places. I do think decades is rightly considered short term for city history, though. At various periods in Paris, it's been laid to siege by viking hordes, had its population cut in half by Black Plaque, been occupied by Nazis, the student protests of the 1960s. Still seems to have done all right over the long run of centuries. Doesn't do much for individuals with bounded lifespans and I know that's mostly what we care about as individuals, but the world and arc of history are a lot bigger than us. None of us will ever know what becomes of Detroit in the 22nd century (unless there are 5 year-olds reading this thread), but there are plenty of reasons it may end up a booming metropolis that makes something other than automobiles and the greatest R&B ever put to wax.

1h agoHN ↗

Amid this uncertainty, there are at least a few reasons for hope that society can better manage a disruptive technological shock this time around. In the past, policymakers were limited by weaker and less knowledgeable institutions than we have today. Social science has made strides in understanding how technological shocks unfold and how to measure the impact of new innovations on labor markets. We also know more about the toolkit governments have to implement policies and how to scale them quickly.

We as in "humanity as a whole" might have the knowledge and the tools... but only in theory.

In practice, most Western governments are falling to the far-right / ultra-libertarian populist crowd. They will not apply what should be done (tax the rich profiting off of AI and redistribute that) simply due to it running contrary to their ideology. And on top of that, most Western countries never came off of the debt they took on during 2007ff, instead adding much more debt during and after Covid and now the Russian war against Ukraine, and probably at least some will also finance urgent aids for the people due to the Iran War with debt.

1h agoHN ↗

What about socializing the means of production???

1h agoHN ↗

Tricky when the super intelligent robots that'll make stuff don't exist yet.

1h agoHN ↗

I think this is a very interesting topic to discuss, and its a shame there is not more detail about how each policy is implemented, or how they might interact with the many other compounding issues we are dealing with.

I would love an opportunity to take my pet ideas and tune them for better scores. To take the data and run it through the AI again over and over.

I would also like to tune the scenario to consider the consequences of thousands of competing AGI's, getting serious about fixing the climate, and the danger of idle hands.