Anthropic et OpenAI baissent leurs prix et, franchement, ce n’est pas une bonne nouvelle.

Ce n’est pas une preuve en soi, mais c’est un signal : si ces baisses ne viennent pas seulement de gains d’efficacité, elles montrent que la valeur créée ne suffit peut-être pas à soutenir les prix attendus. Même avec un marché concentré entre deux ou trois gros acteurs, les investissements restent difficiles à rentabiliser, alors que ces entreprises perdent déjà des milliards.

La bulle dépasse tout ce qui a existé par le passé, et on en est encore là.

Non, franchement, ce n’est pas une bonne nouvelle. On continue de courir vers le mur.

The most innovative and productive companies I’ve worked with had something in common: employees used the company’s products because they wanted to.

That willingness to use the product tells me more about a company than the size of its budget.

A company can hire excellent people and have a convincing plan for the future. But neither guarantees that anyone there will choose the product when they have a choice.

So when I think about what helps a company come up with useful improvements, one of the first things I’d look at is whether its employees choose to use what they’re building.

Considering that veganism is one of the most ethical and sustainable ways of sharing this planet with other living beings

All we need to do is look at what food can be eaten by most people.

If you’re a bystander, occasionally ask about vegan options. This helps: It raises awareness that there are people who do not want anyone to suffer or die for them, and it may prompt places to look into providing such options.

I spend a lot of time fighting against social injustice. It’s important to fight back against bad things, like it’s important to remove weeds. But it’s just as important—maybe even more so—to plant the seeds you do want to grow, so they can crowd out the shit you don’t want.

In native gardening, the most effective way to prevent unwanted plants (aka weeds) from growing is to crowd them out with plants you do want.

As Dungeon Crawler Carl always says, “you can’t save everyone.” But the seeds you plant now spread, faster and farther than you might imagine.

It's difficult to transition from beginner to expert. The documentation are often not meant for the mid, and the export often forget the confusion they experienced as beginners.

The author shows that writing about problems you struggled to solve (especially failures and confusing documentation) helps others find answers and helps experts identify gaps in their explanations. Even if the solution seems trivial in hindsight, the difficulty of finding it is valuable information: if one person struggled, many others likely will too.

L'IA a aidé à maintenir Node avec FreeBSD pour l'instance Peertube.

Est-ce que la documentation disparaît (ou le partage) aussi à cause de l'IA?

Je constate que les personnes s'écharpent sur le sujet, alors que je pensais que Pierre Beyssac était réfléchi et propre dans ses propos, mais je constate que l'IA provoque cet effet whaouh. C'est toujours le libre qui en prends un coup lorsque ses défenseurs passent leur temps à se quereller, ou à soutenir des technologies qui le détruisent.

:focus-visible is why your buttons should NEVER outline: none again.
⌨️ shows the ring for keyboard users
🖱️ hides it on mouse click — the thing designers hated
♿️ you get a11y AND clean visuals, no JS heuristics
🎯 pair with :focus-within to highlight a whole form group

The "ugly outline" problem was solved years ago. Stop removing focus.

Teenagers have started calling AI art "boomer art" and consider it cringe, and YouTubers have stopped using AI-generated thumbnails because teenagers find them cringe and won't click on them. I honestly couldn't be happier.

"“AI is cringe” is one of the better developments to happen this year."

La quasi-totalité des antivirus pour Windows sont désormais des scarewares : Leur rentrée d'argent dépend du fait que vous ayez peur. Donc ils vous font peur afin de vous inciter à payer, avec des messages anxiogènes, quand ils ne sont pas carrément mensongers. Ça s'appelle une arnaque.

This is emotional labor. I don't think it's a coincidence that the overwhelmingly male tech industry has deemed this the work to surrender to machines. The dream of the personal assistant/wife/mother-as-a-service is alive and well in the terminally techbrained. To them, care and empathy are inconveniences to be automated away. Such a deranged set of priorities only makes sense in a world where business output is the highest moral good.

Go is simpler but has a light overhead (2KiB per goroutine and 30 to 75 nanoseconds to spawn one), but Rust has a fragmented ecosystem where async is dependent of the runtime and the related functions comletely differ from the synchronous ones.

Building websites by hands mean the solutions to errors are within you somewhere. "You’re also rewarded with greater understanding of the problem and solution, knowledge that you’ll carry with you."

An AI agent hides, obscures and obfuscate the result or the method used to reach such result.

The web can florish with little websites everywhere. It's never about the size of the audience for the OG web, but always about the community.

Quelques problèmes des affiches générées par IA:

• toutes les informations sont mise au même niveau
• toutes les affiches s'uniformisent. Comment se démarquer des autres ?

Ce style explose, tout comme les affiches de type PowerPoint ou Word s'étaient démocratisée auparavant.

Someone is getting a broader idea that the rules of the web are powerful

The project built without JS is https://textlog.cc/hot

textlog is a simple social text log: write short notes, follow people and hashtags, and join conversations without turning every thought into a performance.
Notes are limited to 500 characters. The constraint keeps them quick to write and read, making room for one thought at a time.

textlog is built around words: notes, people, hashtags, and conversations. It is intentionally small, straightforward, and easy to follow. There are no engagement tricks or pressure to build an audience.

Some thoughts about the AI watermarking:

1. the way it's implemented makes it hardly possible to verify the watermark independently of the LLM's owner.
2. It helps the monster to differentiate fresh food from its own vomit, so it will not feed from the low-quality slop it once created.
3. Something that is not immediately visible as a human, and easily removed automatically is basically the opposite of a watermark.