Very sobering opinion piece. For all the talks about a China / USA race, it feels more like two flavors of the same dystopia. The race is just here to justify acting against their own population interest. The result is then the increase in illiberal fixations and nihilistic world views. This can't end well.

This is an interesting proposal, let's hope it gets picked up and appear in more licenses.

A very balanced set of recommendations from the SFC around LLM uses. It's just the beginning and still lacks a bit in details. It's very welcome though and I look forward to their updates.

There's really something nasty at play. Those coding agents are clearly not insulated from the system enough and to easy to manipulate to exfiltrate sensitive information.

Indeed, if the rsync maintainer can't handle a coding assistant properly... who can?

Very interesting take. This gives very valid ground on why tech communities should reject AI based contributions. Not doing so will indeed hinder the commons communities rely on to exist and improve. This is a path to prevent getting better at inclusivity and diversity (which is really needed).

Interesting point... Didn't think about it this way. We'll see I guess. Maybe human made services will actually get a premium rate indeed. Wouldn't be a bad outcome I guess?

Or why most of the studies we see out there can't be trusted. They're full of holes and flaws. We'd really know people who know what they do in humanities to conduct such studies to get a chance at a proper picture.

Good overview of why we don't see a speed up in development processes when AI tools are introduced. The bottlenecks don't magically get destroyed.

This feels a bit too realistic for my taste... and yet... Well this piece of satire is well crafted I'd say.

It totally makes sense. If you're a FOSS project you have to invest in getting more long term contributors, which requires mentoring. The contributions themselves are not something to maximise. I wish more communities would follow that path.

This part of the industry is struggling more and more (or more likely silently taking more risks to hide the struggle). It has no path to sustainability and it starts to show.

More in depth look at the launch white paper and the issues covered in the PR. Not much survives scrutiny... there's nothing special with this model.

Indeed, and it's going to get even crazier at some point. I guess somewhat soon but who knows...

Are we surprised it's mostly a PR stunt? Not at all. Of course, I agree a lot with the conclusion: we can't trust any claim from those companies. They try to present themselves as labs but mostly try to disguise marketing as research...

It's first a great marketing stunt. The model is likely not the secret sauce though.

Long but very precise piece about why you can likely ignore LLM for development purpose. Starting from older Fred Brooks work is spot on. Indeed whatever will remain of LLM based tools in the years to come, it's much smarter to focus on fundamental skills than chase the new tools. At least, I'm trying to do my share in getting myself and others better at the craft.

Definitely interesting approach. I think neurosymbolic approaches are what we ultimately need so I'm probably biased. At least it means using LLMs for what they're good at (language skills) and only that. Then rely on proper code symbolic models which do the reasoning heavy lifting. I'd expect it can give nice output with smaller models.

Excellent piece, it show quite well the problem of skipping the "grunt work". Without it you can't really learn your trade (be it astrophysics or anything else). It also shows how the incentives on scientific careers are wrong. It's not new, but when LLM agents become available, things are definitely changing for the worst.

So much this... There are so many organisational problems that churning code faster is likely not what you need. When did we start to obsess with the number of lines of code?