Mistral is still a lot of fun, especially with the ChatGPT/Claude voice becoming too common. Openrouter gives you access to it for free, or you can download it to LM Studio.
It says 3 Mar 2026. No downtime recorded on this day. Related: Elevated errors in claude.ai, cowork, platform, claude code
It's definitely not "elevated errors". It's down. Inaccessible. Not green. Yellow at best.
What's extra annoying is it works on some accounts and not on others. It works on Cursor running Claude and Antigravity running Claude, but not on the Claude account I paid for personally.
I think it's only down for some people, and/or some times, and therefore is just having "elevated errors" by their definition. I think "down" really means down for everyone.
A lot of people predicted all this. Check out Ray Dalio's analysis.
Empires rise because of resourcefulness, education, good work ethic and humanity, effective resource allocation (low corruption), high productivity. There's a snowball effect here. These countries become stable. Stability turns them into financial centers.
Then they get rich. There is no work ethic - cultural values reward consumption and not working, which tends to lead into things like colonization and slavery. The wealth gap increases. Wars overextend and become more cost than income.
To catch up with lowering productivity and the habit of war for wealth, there's more debts. They hit a point where it's not possible to pay debts, so printing money becomes excessive. Excessive printing causes them to be dropped as a reserve currency. The wealth gap causes internal conflict. There's more gatekeeping. Castes grow stronger (in modern times, it could be ivy league degrees, gender, and skin color).
All these conflicts and problems end up with weak leaders taking advantage of the situation. Eventually, it ends up in civil war or revolution/reformation.
That's the pattern of rise and decline. It's not chaotic but rather quite predictable.
Instability and poverty is the norm globally; the average person does not have disposable income. While Americans can afford to buy a car with cash lol. It's the US falling to the situation where they're similar to the rest of world. And place like China which are starting to feel things like disposable income and 40-hour work weeks.
I guess just adjust your expectations. The mismatch hurts sanity.
I'd say closer to $5000, which is about the cost of a domestic helper for a year or so.
Then again, they don't need sleep, can be jailbroken, they only need closet space, and won't take your money and run. They can do dangerous tasks like fixing the roof too. Menial labor like walking 3 km over to the store to buy me a can of Red Bull. You can have them do pranks at night. They'd probably be valued quite a bit more than a human.
A robot like that would be far more useful to me than my car (I rarely use a car), and I paid 50k for my car. So for me personally, 50k would be a no-brainer. But of course only if it can do the tasks I mentioned well enough.
Haha, my first thought was why didn't OP just read the comments on the article? Didn't occur to me that someone on HN was finding articles through a source other than HN :3
In the spirit of Hacker News, a good way to learn about these tags is prompt injection and jailbreaking Claude.
I'd post a link, but unfortunately many are highly NSFW. Just search for "Claude jailbreak" on reddit or something.
You'll start to see how Claude really thinks. They'll put things in <ethic_reminders>, <cyber_warning> or <ip_reminder>. You could actually even snip these off in an API, overwrite them, or if your prompt-fu is good, convince Claude that these tags are prompt injections. It's also interesting noting how jailbreaking is easier on thinking mode because the jailbreaking prompts will gaslight Claude into thinking that these tags are attacks.
There's a lot of speculation in this thread, but go and have a spar with Claude instead.
There was an early era when the AI revolution was led by us Java users who had to deal with boilerplate lol. Plus review 4000 lines of code a day, a lot of it boilerplate as well. This was also the era of "clean code" and we were looking at adapters connected to adapters for no reason besides DRY being a holy rule.
reply