It might be if all you're seeking is large-cap stocks with lots of volatility you can leverage that are here to stay for the long haul. Also, the market doesn't seem to believe that Trump will be in power forever.
> "No way the US is going to nationalize a tech company regardless of what happens. The exodus of capital would be unimaginable."
You simply cannot apply any sort of actual logic to the reasoning of the current U.S. government's actions... They just "do stuff" because they feel like it, with no clear thought whatsoever of any potential consequences that may occur.
The CEO of Anthropic himself has said AI is like a nuclear bomb when justifying export controls on Nvidia chips. How many private companies control nuclear bombs?
Why would you think nationalising is a violent process?
As soon as the nation owns enough stock to profit from government decisions (and to compound the influence of those decisions) you essentially have a partially nationalised business.
10% of OpenAI might easily be enough to reach a meaningful "partially nationalised" threshold, once you factor in any holdings in federal pension plans and the active level of government policymaking.
It is very clear Sam Altman wants this, too, because this whole "take 10%" thing in Trump's mind was his idea back in early 2025, and OpenAI have been following up on it recently.
No they didn’t. After Trump started making noise about their CEO, Lip-bu Tan, being Chinese they then took the shares at a “…discount to the current market price.”[1]
And the money for this _deal_ was primarily from the CHIPS act funds they were already awarded but had not been sent to them yet
> Of the total, $5.7 billion of the government funds will come from grants under the CHIPS Act that had been awarded but not paid, and $3.2 billion will come from separate government awards under a program to make secure chips.[1]
This was at gunpoint from the government’s monopoly on violence.
The government had passed a law appropriating funds to subsidize semiconductor manufacturing in the US and spent some of it buying intel stock. How is that the government seizing Intel at gunpoint? I mean aside from the libertarian argument that the taxation necessary to raise those funds is theft?
Trump has already (with Altman directly egging him on) talked about the US taking a share in (i.e. partially nationalising) the AI companies. Has he not called a meeting about this next week?
Their gap over Chinese models like GLM-5.1 is nowhere near 18 months. In many areas, it’s less than 6 months. The best closed models 18 months ago were worse than Qwen3.6.
It was more like November. But it wasn’t really an inflection point, harnesses got good enough that people started noticing by the holiday break. And I’m not discounting some good ol’ stealth marketing in there as well.
Deepseek feels pretty close to Opus at this point, and it’s certainly useful enough for me to spend $20 on api tokens instead of four Claude max plans….
Have you tried deepseek V4? It costs pennies and is as good as Opus 4.6 (I found 4.7 to be a downgrade, and cancelled my claude subscription before 4.8).
Reasonable-sounding arguments always come up in these debates, but in reality it all comes down to what you’re used to.
Disabling natural scrolling used to be the first thing I did on a new system. Until I once was too lazy to do it, got used to it, and now I can’t imagine ever going back.
So I was a Mac user for years and accepted and adapted to natural scrolling after it appeared as the default in 2011. When I switched back to a Windows laptop for work around 2018, I kept it on natural mode.
But then two years ago I got a desktop computer with an external mouse again and.... natural scrolling doesn't work for me on a physical wheel. With a trackpad, the metaphor is direct, that the page or document is being moved by the motion of your fingers; but with a wheel, I still want to pull it toward me to scroll down, because that feels like rolling the little wheel along the document, or turning it to advance the document beneath, like a printer finishing a page.
Maybe that's all silly, but for me it's natural scrolling on trackpads and conventional scrolling on mice with scrollwheels.
Benchmaxxing isn’t the only problem. Evaluating an intelligence is a task that generally requires at least an equally capable intelligence, if not one of greater capability.
That’s why students are evaluated by teachers with more knowledge and experience than them. It follows that any mechanical evaluation scheme is hopelessly inadequate for measuring the true capabilities of a frontier language model.
> students are evaluated by teachers with more knowledge and experience than them
This starts to break down in college when the professors often at best only slightly ahead. (they have more knowledge and experience - but in a slightly different area and so it isn't relevant to the depth of whatever is under consideration) Grad school is about advancing the state of the art - if you don't know more than your professor you are doing it wrong.
> This starts to break down in college when the professors often at best only slightly ahead. (they have more knowledge and experience - but in a slightly different area and so it isn't relevant to the depth of whatever is under consideration)
I can't speak to the humanities, but this estimation is just not true at most universities in the sciences. (EDIT: As cycomanic emphasizes below (https://news.ycombinator.com/item?id=48477683), the part of the original comment pertaining to graduate education is more reasonable. I am speaking here only of undergraduate education.)
It certainly is true in physics and engineering that a PhD student at least half way through their PhD should know more than there supervisor about their topic (and usually much earlier). Even a Masters thesis project student should understand the intricacies of their project better than their supervisor. I'm speaking as someone who has supervised a significant number of both PhD and Masters students.
The original post said “in college”. It might be true for PhD candidates halfway through their program, but that’s like 0.5% of college students. The vast majority of students are leagues behind their instructors in domain knowledge.
I wouldn't say leagues behind, but otherwise I think we are on the same page, though I guess I worded it wrong. It is common for a couple students in any class to know more than the instructor in some niche part of the field even though the instructor has much more knowledge overall.
Yes, I intentionally left out the next part of the quote about graduate school, since that seems more accurate. I was disputing only the part that I took to be pertaining to undergraduate education. The full quote is:
> This starts to break down in college when the professors often at best only slightly ahead. (they have more knowledge and experience - but in a slightly different area and so it isn't relevant to the depth of whatever is under consideration) Grad school is about advancing the state of the art - if you don't know more than your professor you are doing it wrong.
Ah apologies, that's what I get for skim reading and kneejerk replying. I completely agree with you, undergrads are highly unlikely to know more about a subject than their professor (obviously there can always be exceptions).
A grad student is evaluated by how well they are capable of following scientific procedures, communicated their results and have a sufficiently broad knowledge foundation. All that can easily be verified by a professor in a related field since they are very experienced in all those things. They don't actually need to be experts in the specific narrow topic the student has become the world expert in.
> How is this remotely true. You can have verifiable tasks that you can’t do. Where does this idea come from??
That is what benchmarks and intelligence tests are, which are vulnerable to benchmaxing etc. You wont be able to do this by gut feel though, you can create a personal benchmark though.
But point was that personal judgement of intelligence requires high intelligence. Creating a benchmark doesn't require as much but is more vulnerable.
Yet human judgement isn’t subject to side effects like fluency and persuasiveness? It’s like everyone in this thread dismisses benchmarks and then…describes a crappy benchmark.
Sure you can create a personal benchmark. Who will evaluate it, you? How many tasks will it have? How will you evaluate success? Will you know which model is which or will you be blind? Which one will you do first? Ah right, benchmarking.
Also, benchmaxxing isn’t possible when the benchmark and measurements come after the model is released, right?
> I wonder how "free speech absolutists" defend the idea of people in low-income countries using these platforms to spread outrage simply to make themselves a little money
By recognizing that undesirable uses of free speech are the price society pays for having free speech, and by strongly believing that it is a price worth paying.
Just like 1.3 million global road traffic deaths per year are the price society pays for having cars, and believing that people should still be able to freely own and drive cars doesn’t make someone a “car absolutist”.
The idea that free speech should probably be restricted if it turns out that free speech can lead to unpleasant consequences misses the whole point of free speech – in many cases deliberately, I think.
Free speech absolutists just don't defend their position because it devolves into absurdity immediately. It's just a dogwhistle of the far right or people that haven't put any thought into their beliefs.
It’s interesting how the idea that free speech is too important to sacrifice to any other cause, which was the position of Rousseau and other enlightenment luminaries, has supposedly turned from the foundation of humanism into a “dog whistle”.
The implication that if someone is unwilling to compromise on free speech, they must belong to the far right, is certainly revealing.
> Just like 1.3 million global road traffic deaths per year are the price society pays for having cars, and believing that people should still be able to freely own and drive cars doesn’t make someone a “car absolutist”.
Car traffic is heavily regulated to reduce the harm being done by cars/drivers.
Don't forget that undesirable uses of free speech can be made less effective by more speech - as long as what you desire is actually in the interest of the people you want to influence. Like for example this article.
And of course in this case the root problem is not that people have free speech but that they are financially rewarded for using it in bad ways. Financial models that reward impressions are fundamentally bad for society.
Those who are making the undesirable speech can counter their opponent's speech with more speech of their own - and they can afford to outspeak their opponents at any opportunity because they are paid to speak and their opponents are not.
If SpaceX is indeed overvalued (which I have no opinion on), it above all else demonstrates how unimportant spaceflight still is.
SpaceX has been the only game in town for quite a while now, to such an extent that the intelligence agencies of foreign governments have no alternative but to launch with SpaceX. They now do more orbital launches than the entire rest of the world combined. If they really are worth less than Microsoft, then it seems space just doesn’t matter that much, because SpaceX is space for all practical purposes.
I think this is right. For example, GPS is tremendously useful, but you can use it for free. Similarly, while there are commercial weather services, a lot of people access the ad-supported versions[0] or get it from their government’s weather agency.
[0]: which obviously provide some revenue, but much less than the value provided.
Part of the problem with this argument is that for them to be worth a lot, you need both the amount of stuff launched into space to go up massively and for no one else to be able to build a rocket that's anywhere near as good. Spacex is doing really well now, but if Rocket labs, China, Blue Origin, or anyone else manages to make a rocket that's half as good, spacex needs to maintain a monopoly despite not having anything especially magic.
SpaceX is not being valued based on the space part, it's being valued based on Musk's claim that the market for their AI is the entire GDP of the United States. The space stuff is only estimated as about 1% of their addressable market
The real immediate economic value hasn't really changed for decades. Starlink is the most recent significant innovation in that area, and it's worth less than 3% of Google's revenue.
The hyped up reasons for space mattering are all completely speculative and at the limits, stretch the bounds of a sane person's credulity: asteroid mining - ok, there may be an economic case for that, but it's far future; colonies on the Moon and Mars - there's no economic case for those, it's purely a dubious expenditure of tax dollars that has no meaningful economic purpose. Etc.
So, what makes spaceflight "important" beyond what it's already used for and has been used for, for decades?
Nope, just a specific kind. Those who developed and cultivated only a very specific skill set at the expense of all others.
I used to think being a generalist, and having persued technical roles with a people facing element was to my detriment, but it’s turned out to be the best decision I ever made.
Being a generalist was very useful to me 5 years ago. Now AI models have made everyone a generalist. That wide but not terribly deep skillset was immediately devalued by the AI models.
You can argue that the models fuck up 20 percent of the time, or that they make poor code but there is a massive part for the industry that is totally fine with that and I think people ignore it to their detriment.
For context: I came from hardware, Linux, networking, telecoms, and datacenter infrastructure, not software development. I always wanted to go deep, but in practice my brain dragged me across many instead, which unintentionally made me very broad. I kept ending up in organisations where I was pushed back into such roles because its apparently where my "value" is.
I give that context because unlike a lot of you, I’m not a world class FAANG engineer and never will be. It is from this context all of my thoughts on AI flow. I work with people who are trying to use AI to produce work involving entire markets, roles, skillsets and technologies they don't even know exist, let alone understand.
> I had the opposite thought.
Until I recently got pulled back deep into engineering despite not being hands on for close to a decade, so did I. I was pulled in not because of any pure technical capability but instead because it's been recognised the team requires more. The skills I thought served only to help me stay employed in any role in the most basic roles are increasingly turning out to be things other's do not have and are becoming increasingly important.
These are skills I always assumed crucial to “baseline competency” for everyone, but yet where a significant amount of them do not, and these individuals are now finding themselves in positions where they are less useful than me as a result. Many of them can not simply be acquired from AI either, and require years of active growth and practice.
> Being a generalist was very useful to me 5 years ago. Now AI models have made everyone a generalist.
I think they could, but have not. Not at a scale required for me to have significant concern.
AI works as well as the context you can provide, and you don't know what you don’t know. If the context is shallow, so to will be the output, even when it looks convincing and that “looks convincing” part I believe is the most dangerous part.
As an example; I've been (recently) attached to an engineering team, despite last holding that title pre-2015, after AI assisted work contributed to a multi million dollar contract loss. A customer experienced an outage, it was "fixed" and everyone moved on. A month later another outage occurred of a greater scale. A huge amount of time was wasted doubling down on the original AI finding, because the actual root cause had not been identified or understood, because it had been "fixed". Turns out AI had identified and "fixed" a symptom, not a root cause.
I was able to identify and resolve the real issue because I had wider operational and infrastructure context the team lacked, but the damage was done. Trust was gone, the client lost, and layoffs will follow. Those layoffs will be “because of AI,” but not any "10x'ing" of productivity. Instead it will be because plausible but wrong work made it into production and hid a very real problem as a result.
That’s the issue with AI as I see it now. It generates answers that survive initial scrutiny while completely missing wider context leading to cases where more impactful but hidden problems are introduced.
> That wide but not terribly deep skillset was immediately devalued by the AI models.
Perhaps “generalist” was the wrong word here.
Most "engineers" I have worked with are extremely deep in their area and surprisingly limited outside it. Even with AI, they struggle to move beyond their specialty because they lack broader foundations underneath not just modern infrastructure, but a range of areas equally important to the health of a business. My advantage has never been being the best engineer in the room, I knew early in my career I’d never compete with the engineer who can patch our kernel before upstream does, despite wishing I could.
What ended up mattering instead was becoming the "95% guy" across infrastructure, networking, systems, operations, business, customer success, and people management that allows me to work with people/organisations and ultimately connect dots in a way even the best engineers I have worked with can not. AI can help you develop skills in areas you don't have, but starting with most of it in areas in which people have exactly none, and where people seem extremely resistant to developing it with or without AI, has me significantly further ahead in the curve. Ironically, at least in my experience so far, AI has made that more valuable, not less.
> they make poor code
I consider this to be the least important part. We have testing, review, and process for that.
I believe (and have instructed juniors as such) that the real value of valuable technical people has never been producing rockstar code, or being a clone of Linus. It's in having a deep foundational understanding of the building blocks underpinning the now endless layers of abstraction, understanding consequences, tradeoffs, failure patterns, business impact, customer communication, customer wants and needs, and ultimately I guess to sum it up, organisational reality.
This feels more important than ever when they can generate plausible looking technical output instantly that they may be able to validate, but equally produce plausible output in a huge range of areas they absolutely can not, but for which their successes in code have led them to believe they can. Because they underestimate what they don't know and in fact often assume they know far more than they do with no real basis for such a belief.
On the whole, I think I would end my thoughts like this.
For years I lived with the stress that "rockstar" engineers would lead to me eventually becoming irrelevant, much in the same way I might fear AI. So far, being 95% across customers, leadership, sales, support, engineering, and business strategy without losing the technical depth underneath it has meant this fear was unfounded and in fact put me ahead of them. I believe I am not isolated in this, and that in fact we will see more of it.
All else aside, my roles as of the last decade often require me to be in the room and working with humans. AI has not changed this, and there is no current indication it will. The requirement will be that I continue to remain in the room, only now with AI. This is for many reasons including regulatory, because portions of what I do involve systems that if mishandled could lead to more than just a loss of profit. There may be less of these roles, but as it stands I see nothing to indicate they will not exist.
The major blocker for manual labor automation in that fashion is cheap energy. China is ahead of the pack with the States' weight behind aggressive expansion of solar tech, and still can't do that.
reply