alex_duf

28th of april 2025, isn't this before mozilla added lots of AI feature in their browser?

What does "fully caught up" mean in the context of an ever evolving technology? I think I'm in support of open weight models (though there are safety implications), but these things aren't cheap to train and run. This fact alone gives no incentive for leading labs to release cutting edge open weight models. Why spend the money then give the product for free?

Now if "fully caught up" means today's level of intelligence is available for free in two years, by then that level of intelligence means very little

Same here. I've had two generations of X1, my latest one (laptop, not child) is almost 5 years old and I honestly don't know which hardware to pick next in order to run Linux.

At work I have an M4Max 128G, and it's hard to beat that amount of compute, with that build quality.

side note, but I hate that we've reach the point where we don't know what's written by a human and what's written by an LLM.

That goes for a lot of comments here accusing each other of being a bot.

I feel like we've known internet trust at its highest and it can only go one way now.

It's a possibly a managerial thing, which KPI are you improving when spending engineering time on adding IPv6 support?

That said, for their HTTP stack they use fastly (as far as I understand), which should make the shift moderately easier.

Claude uses the wrong modality to be a piloting model. Latency is critical, and outputting tokens in the hope they take the action at the right time is kinda bonkers.

You'd want all the data from the plane to be input neurons, and all the actions to be output neurons.

timing wise this seems to match the Claude Mythos story.

So maybe they're trying to free-up some GPU capacity to run audit of projects in need? I'm assuming Mythos is not cheap to run.

The cache TTL story is also probably link to the RAM price going up like mad so they're trying to save on future expenditure here maybe?

I do understand why people are pissed though

At the rate things are going, yeah there will be a point where intelligence services aren't going to be happily cooperating across the Atlantic, that is unless there is a clear mutual benefit in doing so.

The one EU service I can't find is an easy OIDC provider, like okta or auth0, to manage the users of a SAAS.

Any recommendations?

it's explained in the video, and there's no way I'll be explaining it better than her

I guess it depends on how you frame it. "I've just posted this comment, what do you think" vs "Someone online has just posted this comment, what do you think".

But it does require to know the bias that LLMs have ahead of testing this.

One was taken during daylight on film, which needs to be processed and scanned, the other one was taken at high ISO during night time on a digital camera.

So much interpretation is done on colour on each step of the way that it's not surprising the colours are looking different.

Or maybe Microsoft continuously degrading their user's experiences has finally reached a breaking point, to the point where Pewdipie AND Linus Tech Tips are talking more seriously about linux.

So yeah Pewdiepie is part of it, but I honestly think that's only because Microsoft has done such a poor job at maintaining their operating system.

corporate speech existed long before AI

everyone here is commenting how odd it looks to use a regexp for sentiment analysis, but it depends what they're trying to do.

It could be used as a feedback when they do A/B test and they can compare which version of the model is getting more insult than the other. It doesn't matter if the list is exhaustive or even sane, what matters is how you compare it to the other.

Perfect? no. Good and cheap indicator? maybe.

The same way you can decide to not visit a website, you can decide to not use LibreOffice.

I think it makes sense, how else are they supposed to fundraise and develop the features that makes the software useful to manage your own data?

right? windows 98 ergonomics (that's NOT sarcasm) with more customisation and not trying to re-invent the wheel every month.

I'm a gnome user myself, I do enjoy the year on year refinement they have going there, but I completely understand and respect the XFCE crowd.

I've installed linux (debian LTS with XFCE) on my mom's computer and she recently called me to thank me. She says her computer is much quieter now (meaning fewer notifications). She only needs a web browser and a text editor.

So you're right, it's great for power users, it's also great for other users.

have you thought about switching to another OS?

I use claude code with an API key and pay per token, and the /cost command is very helpful.

And before people ask, it's because I have a very low usage and it's cheaper to pay per token. I'll have the odd month at $30, then nothing for a few months

Do you mean every pixel or every sub-pixel? Sub-pixel is interesting because the geometry of the grid isn't going to be the same from one screen to the other. It might also look compressed horizontally.

This claim keeps circulating around and around and is not "EVs are producing more pollution", it's "if EVs are going at motorway speed, and if we only look at the pollution generated by the tires, then indeed they produce a little more".

But that's completely ignoring tailpipe emission, and the fact that in an urban setting it's still vastly more advantageous to drive an EV.

See https://doi.org/10.1016/j.trd.2025.104622

I don't think it's about repeating the instructions, but rather providing feedback as to why it's not working.

I've noticed the same thing when creating an agentic loop, if the model outputs a syntax error, just automatically feed it back to the LLM and give it a second chance. It dramatically increases the success rate.

Exactly, I stopped reading when I saw the flying stats. There are people who still haven't clocked where our climate is headed.

I get that you may have to see family abroad or maybe indulge for a holiday, but this is "I'm using an airplane to commute" kind of level.

And here I am trying to book my train tickets to go to London instead of flying even though it costs three times as much just to avoid a few kg of CO2 (among other things), it's making me angry.

In fairness, back in 2017 I bought a OnePlus 5T with 12G of RAM.

That's almost a decade ago.

Phones RAM progression has stagnated for a LONG time, during that time I doubt that webpages have become lighter, so yeah I'm not surprised by what you are saying.

A lot of hate in the comments, I think it's great that companies are in a position where they think it makes sense financially to support Linux as a target platform.

If I may give constructing feedback then, "noise evidence generator" sounds like you're creating fake noise proof to fool courts. The name needs to be revised.

"recorder" might feel less like it's inventing things.

For grid storage the only two things that matter are cost and reliability. So it would be a big deal, but it's not the public for whom it would be the biggest deal.

I think light aircraft doing short flights might be pretty interested by the technology. I remember 400 Wh being a threshold above which flights become feasible.

Starting his post by "I hate corporations and Google is yet another disgusting example of it" is a great way to have no issue resolution. Assuming good intention (at least in writing) may help if this post gets traction.

Politically it would be an interesting choice for Netflix to encourage people to use their BitTorrent clients..

But technically, you're right.