xienze

> The United States was a couple feet from an actual coup dude.

How does this square with the oft-repeated line that the Second Amendment is useless when the government has tanks, bombs, and aircraft carriers? January 6th was a bunch of unarmed boomers. If you really think the government was >this< close to being overthrown by them, then you might as well pack it up. The US wouldn't stand a chance if a couple thousand armed Russian or Chinese nationals decided to show up to the Capitol.

Well they do have an IPO coming up.

> but he chose to do it with a Chinese spy

The implication here being that he _knew_ she was a Chinese spy.

Weird, I've been told repeatedly that inference is wildly profitable for both of these companies and waves hands training costs are irrelevant for some reason.

> The Asahi Linux project kicked off in December 2020. About 2 years before the first ChatGPT (Nov 2022). It had a lot of ideals including working with the BSDs at the time, and so they surveyed the landscape and considered the downsides at the time vs the upsides and decided a BSD approach was the right tradeoff. Now, 6 years later, the landscape has changed pretty radically

Asahi themselves cited WINE forks from the early 2000's as a case study in "bad actors" abusing permissive licenses. So AI really has nothing to do with this, they were well aware of the precedent in this area.

To me it's just another example of this weird hypocritical stance some free software people have. They love putting the code out there and letting it spread far and wide, but once someone _else_ starts making money building off it, "hey no fair!"

Who said they didn't? Not sure what "very little" attribution means, but presumably it doesn't mean "none."

Maybe what we need is better messaging from the people building data centers.

WHY do we need infinity data centers? To "win" the AI race? Tell us in quantifiable terms what constitutes "winning."

And we already have 10x the data centers that China has. Can't we pump the brakes a little bit? If not, tell us why not without using vague language.

Finally, tell us, again in quantifiable terms, how building a data center will benefit us. Tell us exactly how many PERMANENT jobs are going to be created, how much tax revenue is going to be created, and let us hold you to it. There HAS to be something in it for us.

Aren't they still losing billions per year?

There's a little more to it than "just a launcher." My Steam account dates back to the release of Half-Life 2. Loads of games, thousands of dollars spent, and a service that's been reliable for over 20 years. Why would I use a different service? I'm not unique, everyone else using Steam is in the same boat, just with less time invested.

They've undoubtedly locked us all in, but they've shown themselves to be trustworthy and reliable.

> How much of this is the current uncertainty over politics in the US?

Uncertainty that just developed overnight?

> With the incredibly frustrating aspect that we can't really directly discuss exactly how corrupt a few of the main actors have been, to date.

What does that have to do with Intel, Amazon, and Google's financial results?

And where did you get the impression that no one can "discuss" any of this? A minute doesn't go by that someone isn't talking about Trump. Including you, regarding a completely unrelated topic.

Surely you agree that "both sides" think the other side is full of evil politicians that want to destroy the country AND their voters are complete morons?

And surely you can agree that each side has key issues they feel very strongly about (abortion on the left, guns on the right) and again, think the other side's view on said issues are utterly disgusting, right?

And SURELY you can agree that gerrymandering has occurred at various times in the past, from both sides, can't you? For example, take a look at the map of Illinois congressional districts[0]. Perfectly sensible districts here, yessir!

The low IQ take is pretending there aren't places where both sides do in fact behave the same way.

0: https://en.wikipedia.org/wiki/Illinois%27s_congressional_dis...

How did they survive before it?

Being sarcastic here. The usual argument everyone brings up "inference is extremely profitable, checkmate AI skeptics!" while dismissing the black hole that is training.

It's very simple, if you just assume they never have to spend money on training ever again, they're wildly profitable.

The closed models are on Openrouter as well. But to the original point, you're not likely to see many people use Openrouter to run Claude or Codex, so it's not good for making comparisons on that basis.

The other keys were 56, 66, 115, 155, 156, 230, 237. I honestly think whoever did this thought the "384" in 384-bit meant they had to choose numbers between 0-384.

> You do know that capitalism requires labor and a shrinking workforce is undesirable for capital owners, right?

Ah, is that why the "smash capitalism" crowd rallies around reduced immigration?

> you could just deport them if they misbehave

LOL, yeah that's gone over real well in the US.

> is a policy decision pushed by climate activists?

No, but there is a significant overlap between climate activists and "we need more immigrants, and definitely leave those illegal ones alone" activists. And more to your point, have there BEEN any climate activists sounding the alarm over mass migration?

You're making the mistake of not looking down the road far enough. "China's developing rapidly and has all these cool EVs, looks like smooth sailing til the end of time for them!" Their male/female imbalance and TFR < 1 are absolutely massive looming issues. Maybe they'll advance to the "robots do everything and no one ever has to work" end-game and then those things won't really matter, but I wouldn't bet on it.

> searxng gets rate limited fast and the results are inconsistent depending on what providers it can reach at any given time / isn't captcha'd out.

Exactly. I really wonder what the people going on about using SearXNG are actually doing, because in practice I've seen just what you see. Even from a residential IP, you might get a handful of Google searches before getting backed off. If you ever have multiple searches going on at once, even Bing and Brave are going to back you off pretty quick. My home IP is apparently permanently banned from using Brave Search through SearXNG.

And when you aren't getting results from Google/Bing/Brave, it gets bad really quick for your agent. As much as I wanted to go fully self-hosted for agent search, it just doesn't work well.

> Clearly a future where all of the "thinking problems" are easily solved but we still have to pick up trash kinda sucks.

Oh don't worry Mr. Former Software Engineer. You won't be picking up the trash or cleaning toilets. You have things like an expectation for your work to pay a reasonable wage. You'll be passed over for a third world immigrant who'll work for a pittance, and they'll be replaced only when robots are capable enough and cheap enough to undercut _them_.

I'd heard an interesting theory that in training on customer chats there's a lot of interactions like this:

Agent: _wall of text_

User: yeah sounds great, do it

Which has the predictable outcome of biasing towards wall of text responses.

> Half of Europeans vote right.

<insert "European right-wing is actually further left than America's left wing" comment here>

> Excel, VSCode, Outlook, etc the permission model is a modal, "Do you trust this?" binary choice to enable full permissions to everything.

Because years of experience has shown that even users who should know better (developers) get extreme permission fatigue. The demand end users make is "it should just know without me having to click 'approve' all the time." Well, that's tough. You're basically stuck between a rock (fine-grained permissions) and a hard place (don't make me deal with it).

> It likely does not fully cover amortized R&D just like the other reply stated but it does still cover marginal inference cost (GPU/power/etc)

Why does this point come up over and over again, pretending that you can truly separate training and inference costs. Yes, they are separate things but the value OpenAI and Anthropic are presenting to the world is "we're the absolute best, no one else comes close." Well, to keep that up you can't just not train for extended periods of time. You have to keep that engine going non-stop when there's free Chinese models nipping at your heels. You can be profitable on "just inference" all you want but if training expenses dwarf that, you're not going to be profitable overall, and that's the bottom line.

> Moviepass failed because they thought it's going to be like Gym membership - people buy and don't go, but guess what? People love going to movie theaters.

You think developers won't/don't stretch subscriptions to the absolute limit? The AI subscription model is like the gym model except a large percentage of the customers work out 24/7/365.

> I think we're going to suddenly see them greatly scale back training and try to sell inference-only

Remember a few weeks ago when all the AI labs said "we need to slow down, to uh, prevent destroying the world"?

> You can paste in pngs and they'll convert them to svg with varying degrees of success.

I've found Opus to be REALLY good at converting images that are well-suited to being SVGs (decent resolution, sharp edges, clear color boundaries [though gradients do OK]) after a few rounds of back-and-forth. It'll do precise measurements to figure out curvature, the exact colors to use/gradient stepping, simplify complex paths, etc.

> Are you concerned about that, too?

Kind of, because every time you stomp out one of Mother Nature's ways to keep the human population in check, you contribute to letting the population exceed carrying capacity, which leads to things like:

> changes humans do to our ecosystems that we know cause harm, but aren't stopping

Bash scripts don't really help when you're dealing with MCP tools or other tools that are local to the harness itself.

Pi doesn't even support a permissions model. It's extremely barebones, you're expected to customize basically everything.