sippeangelo

That's why I'm so puzzled to why Composer doesn't work better when they have the ability to train it from scratch for their agent harness! Yet it still fails to apply edits, gets confused why it can't call some commands in its sandbox, the list goes on...

It definitely feels sufficient for questions and planning, but it is surprisingly lacking in the actual coding department once you go for edits that need changes in multiple files. Which is surprising considering they should have been able to train it on their own harness!

That's a hefty payday for a model that barely functions! Every time I run out of API credits and get kicked back to Composer 2 I feel like I'm better off just packing up for the rest of the month.

I feel like we're finally at a point where you don't have to constantly argue with and constantly babysit coding models, which makes it even more frustrating when you're suddenly forced to deal with one that ignores your instructions and gets stuck in thinking loops again.

I suspect it's the vast troves of training data rather than any tech that Cursor possesses that SpaceX is after...

Crypto has been an awful development in many ways, but I happily welcome it when it has made malware so much more benign to me. The last malware that affected me personally was a crypto miner worm, and the one before that was a crypto wallet stealer, neither of which affects me at all as I don't meddle with crypto.

I don't know the statistics, but it seems like it's way more profitable for the grifters to target other grifters instead of taking over my machines and extorting me. Or maybe I just got lucky.

Then why 8 instead of 3?

Then maybe it shouldn't be done? What??

Some Chinese sources sell modded Nvidia GPUs with extra VRAM. They're quite affordable in comparison to even a Mac Pro.

I went on to install this, but it seems very US centric, which isn't apparent in anything else than the domain name. The maps only cover the US, you can only download English dumps of Wikipedia, etc.

It's not the biggest deal if you're proficient in English, but I wasn't even able to download the full dump of English Wikipedia as their hardcoded link to it just seems to return 404.

The Docker setup leaves much to be desired, as network names are hardcoded, and extension services are expected to be reachable over hardcoded port numbers, making it impossible to run behind a reverse proxy.

Going to give this another go in a couple of years when it has had some more time in the oven, but it still looks very promising!

4chan is clearly operating in the US. The UK can easily cut the overseas cables and fix the problem!

I'm having a hard time understanding what this is. I was hoping it would be tools to manage secrets for AI agents, but it looks more like an "Enterprise trojan" type snitch for when your employees use AI tools and passwords are shared?

Said scoreboard: https://www.project2025.observer

Please explain how opposition to privacy invasive solutions result in even more privacy invasive solutions being implemented? Is it purely out of spite from the lawmakers? This logic doesn't follow.

Microwaves are very directional. What about it sounds implausible?

From their video it just comes across as they stimulate different left/right neurons depending on where the enemy is on screen and then listen to some output that also says left/right. Shooting looks completely random, to be frank.

If you connected electrodes to two different fish, shocked them and interpreted twitching as intelligent output, fish could also play Doom. The interface is doing all the work.

It doesn't sound like the neurons have any concept of the game other than "left input means left output", which is a rather trivial result... It's effectively no different than the pong example.

They don't say anything on how much training is required for this to happen, or if there's any "learning" going on at all. The learning part is "next".

Maybe stop?

It is actively harmful to society. Slap SynthID on some of the photographic evidence from the unreleased Epstein files and instantly de-legitimize it. Launder a SynthID image through a watermark free model and it's legit again. The fact that it exists at all can't be interpreted in any other way than malice.

Bit of a sensational title? This doesn't "break WiFi encryption", only device isolation if the attacker is already in the same network.

.NET Framework 5 or .NET Core 5?

Yeah this is truly bizarre. I've been assuming that their "auto" mode basically chooses their own Composer model when you run out of API tokens, but now their homebrew model is supposedly more expensive than the mainstream offerings? How does this reflect on my Cursor subscription? Why is it more expensive in the first place when it's "just" Composer 1 + more reinforcement learning on top? How can it be more expensive to run? Or have they just decided to stop subsidizing it?

"Tera" always sounded cool though.

I can't imagine a world where I had to work with the Gitlab CI pipeline on the daily. I'm sure it's not perfect, but what are you even doing if you have to touch it DAILY?!

Yeah, that's what I'm trying to say! The furthest right icon of this post is peak skeuomorphism, but we've never actually gotten back to it. Someone's always gone "wait, this icon looks a bit too much like the real thing, we can't have that!". It has never been cyclical!

Is it? I've seen us going from obvious skeuomorphism to more and more abstract shapes, until we hit peak Windows 8 hubris where everything is a coloured square with a monochrome symbol in it. Then back to icons where shades of colour and contrast finally start meaning things again, but getting stuck in an endless balancing on the edge where icons are abstract enough to confuse but not clear enough to describe their function. We've never gotten fully back to actual skeuomorphism.

Governments using Palantir services as a loophole to enable mass surveillance by linking data is the evil part.

That's not a turn-taking model, it's just a silence detection Python script based on whatever text comes out of Whisper...

With all respect to Mozilla, "respects robots.txt" makes this effectively DoA. AI agents are a form of user agent like any other when initiated by a human, no matter the personal opinion of the content publisher (unlike the egregious automated /scraping/ done for model training).

To be clear, this is Reddit auto-translating and poisoning their own search results, and it's the most frustrating thing. Post AI and post Google enshittification, we all ended up appending "reddit" to the end of our searches to filter out all the blogspam and get some real human opinions on queries, so Reddit decides to cash in and make sure they receive the maximum amount of that traffic by making it available in every language under the sun, poisoning their own well deliberately. It's shit turtles all the way down!

Thankfully the feature is just in time for it to fall out of fashion! It really is an awful layout, UX wise. But at least it looks pretty at a glance!

You use Syncthing for object storage?

The overmoulding is seriously ugly. Gold and navy blue?! Silver and medical grey? Only the black is passable.