18,091 karma · joined January 24, 2012
Or are they just reselling access to some ahared legacy infrastructure? In which case any "competition" is very likely an illusion.
AI isn't like any other historical tool or technology. Its qualitatively different. Flint axes or waterwheels or a plasma torches dont just suddenly appear everywhere, and then start getting rapidly (and likely recursively) better at the thing they do. While also mimicking innate human characteristics like complex language.
Its the rate of change.
So two questions:
1. TFA mentions the "humourous phrase hash" theory, but Claude's trajectory seems to be a long way from that. To what extent was Claude prompted in the ultimately successful direction?
2. Whats with all the "hmm..." in the internal reasoning process log? Do models really do that? And if so, why? Some weird emergent language thing, or just mimicry?
Like the Cybertruck, this is just the Musk "fuck you, get out of my way" worldview realised as metal and plastic.
As for "small island" - I was trying to make a point about the size of the uk on a global scale, and compared to vast places like the Argentinian campo. But perhaps "island group" or "archipelago" would have been a better choice.
In the 70s and 80s (and probably before) in the UK, the BBC used to broadcast occasional "sos messages" [1] prior to the main daily news summaries on FM radio - for people who were away from home and, for whatever reason, uncontactable. They usually related to a relative being dangerously ill. I remember hearing them on my parent's car radio on summer camping trips.
Obviously the uk is a small and quite densely populated island but, without two way personal communications, some needs were similar to those described in the article. It changed a lot of things.
[1] https://swling.com/blog/2012/09/radio-4-and-now-an-urgent-so...
Seems like this might be another example of The Bitter Lesson [1]:
"The bitter lesson is the observation in artificial intelligence that, in the long run, general approaches that scale with available computational power tend to outperform ones based on domain-specific understanding because they are better at taking advantage of the falling cost of computation over time."
AIs don't need comments that are designed to work around human cognitive limitations. They're quite happy to just crunch the code and extract the meaning.
Now in 2026 the main generators are EDF (French state owned), RWE (German), and Scottish Power (Spanish owned). These companies tend not to publish executive pay, but back in 2012 the CEO of EDF in the UK was being paid £1.2 million a year.
I would submit that all this has some bearing on why the prices have gone up.
By "here" I meant posts on this thread, not in the content of the article.
I also wonder what the contribution of domestic solar and on-site industrial generation is? I'm a dev in a large manufacturing business and two of our three factory buildings are now covered in solar cells which power a lot of our operations. Some big industrial sites have on-site wind-powered generation too. Is that counted as generation in these stats? Or is it just manifested as reduced demand?
(Edit: added "being posted", "welcome" and "(neutral?)" for clarity)
And to preempt the inevitable comments about what happens to the employees: anyone who works for an oil or gas (or tobacco or social media or junk food, etc) company who's culture is so broken that it produces such planet- or society-wrecking egregious behaviour, is part of the problem and they know it. Find a new job doing something constructive.
Handyman? One of my neighbours used to be a senior sales director. Now he paints fences and assembles flatpack furniture. He seems happy enough.
Or get out of town, buy a shack with some land, try live cheapy off grid, wreck my body hacking turnips from the frozen waterlogged ground.
Retire, try to live off an inadequate pension, hang out in the local public library where its usually warm.
I've long suspected the RFK jr's role is simply to exploit that for his, and his political owner's, benefit. Thata all. The man has no redeeming qualities.
This seems like a rather fascile take, tbh. Along the lines of "everything is just atoms". Yes they're numbers. And yes, the hardware is just atoms. But its the emergent properties of those things that is significant, and the way that behaviour is embedded in human society - which can be just as morally consequential as swords and brains and harmful ideas.
The Orange Entity sets the direction and his congress provides the money. My guess would be that NASA has been told that the direction that leads to there being money is to forget the space telescopes and other science, and instead to focus on national prestige projects (Moon, Mars) and things that will feed through to military applications (hypersonic flight).
Clearly. And I'm sure you're right about US public sentiment.
I was perhaps trying to make the point that powerful authoritarians like Trump always have to consider future threats because, unless they can stay in power forever, the future is where they are most vulnerable. Eventually they no longer have the energy to keep it going, and someone else is in charge - someone who then has more power than they do, and may not owe them any loyalty.
The usual work-around is dynastic power, but I don't see any signs of that this Trump's case.
> But he'd belong in American jail, tried in our courts, under our legal system.
For January 6th, sure. But the war with Iran, the Venezuela thing - they're international crimes the like of which a non-trivial proportion of the world considers the ICC to have jurisdiction over.
But anyway... if he doesn't die in office he'll realistically be too old to face justice in any court.
Personally I'd love to see Trump on trial at the ICC. Along with Putin and Netanyahu. And Blair (the former prime minister of my country) and George Bush.
As for me, I've come to appreciate that I'm a fairly average thinker. Mostly I just try to make things work and watch the clouds.
There are people who's thinking I admire - Stewart Brand, for example. As for Nietzsche et al, I've never really had the time.
I don't know for sure but, as others have pointed out, yesterday's submission was the third time you'd posted it to Show HN. People have a low tolerance for dupe posts, especially if they're perceived as excessive self-promotion.
Also, your work might be seen as vibecoded and therefore "ai slop". I dont know if it is, but all it takes is a perception.
Finally, your work might be seen as a commercial threat by some people here - causing them to flag it.
> how i can share my work with HN community
Write a technical article about some aspect of what you've done. Talk about the challenges and how you solved them. Include a link to the repo. Put it in an external site (even github) and submit a link. If its interesting then people will notice.
Musk would be so proud. Do you by any chance drive a cybertruck too?
Seriously, don't do this while you're driving. Or even while your car is driving. Its dangerous to be distracted in that situation.