Ask HN: Should development of AI be nationalized? | Hacker News Reader

Ask HN: Should development of AI be nationalized?

And…

Should the US insist that all of its allies nationalize development of AI?

For example, and for starters, the US could push for NATO membership to be amended such that all member states nationalize AI and agree to sharing all source code, training data, methodologies and results for research and development, as well details of budgets and expenditures for the same, with allowances for limited (and time-limited) non-disclosure for the sake of security.

It could be up to each nation to decide how to provide their citizens, institutions, and businesses access to models running on public infrastructure; and whether and to what extent private entities may operate and expose models on their private infrastructure.

View on HN

>Should development of AI be nationalized?

Should anything be nationalized? Basically never. At most minimal regulation is needed; not ownership.

Energy and water work better with competition. Most countries moving toward greater privatization.

Healthcare? Even Canada's "public" system is heavily tiered and powered by private big corps.

Telecom? Clearly private. Bell, AT&T... nationalization as terrible.

Journalism? State media is inevitable propaganda.

Military? 100% returns over 5 years of lockheed, nortrop, or boeing.

Infrastructure? Roads are built by private contractors. No one has nationalized paving crews and nobody is considering it.

We don’t nationalize; government can give $, regulate transparency and liability.

>For example, and for starters, the US could push for NATO membership to be amended such that all member states nationalize AI

Do you think the republicans/trump will do this?

> Energy and water work better with competition.

I haven't seen any real evidence of this, but I've certainly seen where privatizing these things made them worse and more expensive.

> Healthcare?

At least in the US, having healthcare be a for-profit thing has been utterly disastrous.

>I haven't seen any real evidence of this, but I've certainly seen where privatizing these things made them worse and more expensive.

You phrase this significantly based on what youve seen.

In the usa, Duke energy, PGE, Con Edison, and countless new solar and wind setups.

For water, yes, still pretty heavily publicly owned. American Water, united utilities, or Suez North America though are exceptions.

This doesnt count culligan and similar, it doesnt count private water dispensaries.

>At least in the US, having healthcare be a for-profit thing has been utterly disastrous.

Big pharma is worth a trillion gdp a year and provides 50% of the world's medicine.

75% of new medicines come from the usa.

90% of bleeding edge medical comes from usa research scientists.

If you want the "best" medical help in the world, you go to the usa.

What you call "utterly disasterous" is more social or the consequences of 74% of USA being overweight or worse.

> You phrase this significantly based on what youve seen.

As do we all.

> countless new solar and wind setups.

I'm speaking about nationalizing existing things, not about new things.

> worth a trillion gdp a year

What they're worth monetarily isn't really relevant to the discussion.

> If you want the "best" medical help in the world, you go to the usa.

If you're wealthy, yes. If you're not, the US isn't a great place to get medical care.

> What you call "utterly disasterous"

What I call "utterly disastrous" is that people die because they can't afford to get adequate health care.

Fwiw, I asked ChatGPT (o3) to speculate on the outcome:

https://chatgpt.com/share/685451c2-8824-8011-800c-49ecd67082...

Google's Gemini (2.5 Flash)

https://g.co/gemini/share/076aa106ceaa

Microsoft Copilot (model ID unknown)

https://copilot.microsoft.com/shares/xLcveZhDMankN6rLin2AV

Research which needs to be a secret already is treated as such, so there’s no real benefit to adding further restrictions.

Trade control is a useful tool, which is also already used where applicable, and researchers have observed that trade restrictions (high wall, deep moat, etc.) don’t necessarily produce the desired effect. For example, see Deepseek and compute restrictions.

In general, science, research, technology, and coding benefits from open-sourced collaboration based on the nature of how activities in these domains are conducted. IMO, scientists and researchers would benefit from worrying less about the distractions of politics, unless they’re explicitly told not to share secrets. Great ideas and work can come from anywhere.

Reply on news.ycombinator.com