I'd say there's also oversight/supervision. Which was manual at the start with a human signing off on commands/incrementally built allow/block lists, and now seperate models evaluating commands and blocking them based on some parameters. This is the weakest model, but it'll evolve as well.
HN user
Narretz
Iirc correctly Opus 4.7 had the same problem, safety filters were triggered way too easily at the beginning.
There a few benchmarks out there where all existing models have abysmal scores. So it's not actually a problem if Antrophic's older models are bad, especially if the jump to the newest model is huge, and the competition is also way below it.
They are now reporting that is a major incident:
We are investigating authentication issues leading to failure in starting Actions runs and downloading actions. At this time the majority of Actions runs is impacted.
Cache entry Linux-pnpm-store-6f9233a50def742c09fde54f56553d6b449a535adf87d4083690539f49ae4da11 (1.1 GB) saved to GitHub Actions cache for TanStack/router, scope refs/heads/main — keyed to match what release.yml will look up on the next push to main
Imo I think this shouldn't have been possible, as in release should use its own cache and rebuild the rest fresh. It's one thing that the main <> fork boundary was breached, but imo the release process should have run fresh without any caches. Of course hindsight is 20/20.
--before doesn't save you globally, only min-release-age does, which is in npm since March iirc.
Isn't this article wrong about npm minumum release age. 1. The config is min-release-age. 2. For some reason they have chosen to make it days instead of minutes: https://docs.npmjs.com/cli/v11/using-npm/config#min-release-...
Completely unforced fragmentation of the dependency manager space imo
Dispatch and scheduled tasks have been available for a few weeks already, although with limitations.
I can't check how much usage I have left because that part of the Claude website does not load ...
Kinda crazy that ts-node is still the recommendation when it hasn't been updated since 2023. And likewise crazy that no other lib has emerged that has typescript compilation and typechecking. Of course if it works, don't fix it, but typescript has evolved quite a bit since 2023.
They are talking about the cost to run the Actions control plane and the scheduler that is not executed on the runner itself.
Amazon Prime did it as well and it was absolutely horrible. Character names got completely butchered, and changed from one scene to the next.
That was 10 years ago. The stability has improved massively. What was the last phone that bent under normal circumstances?
It still has 6.5 inch display and the camera sticks out like a sore thumb. Where's a 5 inch display normal thickness phone?
I have a crazy suggestion ... maybe Congress should make this a law?! Does Congress still do this outside of insane huge budget bills?
Trump's handling of military strikes/operations, which have been mostly symbolic. Killing Soleimani, and not retaliating to the retaliationary strikes. A completely useless strike on Shayrat airbase in Syria. Pulling out of Yemen strikes this year because it was ineffective (never admitting to this though). Trump wants to be known as a deal maker. I don't think that has changed, he's just become more delusional regarding the practicality.
Israel doesn't talk about destroying Gaza, it just does it.
I guess because many men are needed for the IRGC and related organisations.
That's a big leap. Nothing suggests a ground operation or occupation, which was the most costly part of the Iraq war, and importantly, was part of it from the beginning. Experience suggests that Trump would rather walk away from Iran after an exchange of strikes and claim victory then double down in a land war.
There's no way of knowing that Russia wouldn't have incited the "rebellions" anyway. Once the writing was on the wall that the majority of Ukrainians didn't want to be Russia's puppets, Putin would likely have acted one way or the other. Why take chances?
It can't be coincidence that a few weeks ago users wanted to twist grok's arm and make it post right-wing aligned answers/opinions, but grok itself said it's programmed for unbiased/factual answers (for what it's worth). This is probably a test run gone wrong to make grok more aligned with Musk's opinions.
It just "happens" to be also available to select for the end-user after product-launch
RAW (any format) is an essential requirement for many photographers. You just can't get the same results out of a jpeg.
Wdym with "curate your feed"? See only posts of people your follow? That's always been possible with Twitter.
This can easily also be because they don't have enough bandwidth to serve the normal request volume because some load balancer update or similar went wrong.
Well it's always good to be specific, but I think it was the only thing that the US did for the F-16, wasn't it? They didn't exactly support sending planes in the first place. And it's not gonna be the last wrench the US will throw into Ukraine's (and Europe's) gears. It all piles up.
Israel is using these bombs on Hamas and Hezbollah mostly, both of which it hasn't defeated after 1 1/2 years of complete material superiority. It's actually impossible to defeat an enemy that uses asymmetric warfare. Israel already knows this, but Netanjahu needs the war to continue to stay in power.
This is interesting. If the models had enough actual code as training data, that forum post code should have very little weight, shouldn't it? Why do the LLMs prefer it?
Wow, they still render to canvas? That's terrible. Just looked at this website: https://payamzahedi.com/toastification/
Feels janky just when scrolling. This is supposed to be the future?
Last time I used a flutter app they didn't even update the url for different screens.
I don't think the article outright claims Oracle has no JavaScript runtime, only that Oracle JET is no runtime, which is true. And since this is the evidence Oracle presented to keep the trademark, it's fair to point out that this is nonsense. But it's also true that if this goes to court, Oracle could present GraalJS (which is used in OracleDB) as evidence for their case.
Several months later, Google launched the most viral new AI application of the year, called NotebookLM
When I read this I was honestly confused. I had never heard of NotebookLM before.