cebert

I’ll always remember him from Twit. He always made the show entertaining. I wonder if he and Leo were ever able to reconcile their differences.

At least the sponsored posts are currently visibly highlighted. As a user, I would quickly be trained to avoid the first N sponsored suggestions and skip the first non-sponsored suggestion. What's the goal here?

Yes! That’s it. I forgot the name, but it was fun.

When I was growing up, there were two airport simulator games that were pretty fun to play. One of them I remember playing on my Sega Genesis.

I forgot about Stack Overflow. What a blast from the past.

I think this is slightly misleading. Modern OSes intentionally use available RAM for caching and will reclaim it when needed. The issue isn’t that Windows is “using 6.7 GB”, but whether the system is under memory pressure and paging. I’m not convinced the author fully understands that.

I was immediately turned away when the UX looked like pure AI slop.

This is absolutely frightening.

Simon has stated a few times that he knows it’s possible that pelicans could be in the training sets. He also has other tests he doesn’t share publicly. He’s just a fan of pelicans.

No, I am more than capable of reading and reaching my own conclusions.

The central claim in this blog post hinges on there being no human review as part of the Bun rewrite. The actual postmortem describes the verification process in detail. Sumner monitored the workflows for most of the 11 days, manually reading outputs to catch issues and prompting Claude to fix the loop. It had multiple arounds of security reviews etc. I see the framing here a bit ingenuous.

gift link: https://www.wsj.com/lifestyle/careers/ai-knows-what-you-did-...

If we don’t do something about this, I fear future generations will not view us in a positive light.

That’s a great one.

With mass drone surveillance and online safety acts, we will finally be able to keep our children truly safe for the small cost of privacy.

It's interesting that the author didn't mention considering updating their agentic code review prompt to keep an eye out for repetitive/duplicate code.

I'm also noticing people mimicking LLM's speech patterns.

I've noticed this happening to myself. I use Claude Code quite a bit at work, and Claude tends to favor certain phrases, like, "the smoking gun", that I didn't use very often before. After encountering phrases like that frequently, I've found myself incorporating some of them into my own speech.

If we can’t bring more public awareness to this issue to put pressure on political leaders, it’s going to be increasingly impossible to avoid governments keeping the children and us “safe” on the internet.

Anthropic has already stated that they plan to keep Fable accessible in some form on subscription plans, but they’re working the details out.

[source]: https://x.com/trq212/status/2072814903170408784?s=46

They want to hold you captive to subscription plans. If they allow CarPlay, it gives consumers more options for cellular connectivity, music, navigation, and other apps.

Yeah, the gift links seem a bit buggy.

Gift link: https://www.wsj.com/tech/ai/the-anthropic-fable-ban-is-over-...

Is there a bug? I don’t see a single curated example from the list or map that I can click and use?

This seems much less capable than Claude Design or Figma. I don’t see much value here as most LLMs could quickly make these prototypes too.

[dupe] https://news.ycombinator.com/item?id=48747304 743 points, 260 comments 12 hours ago

It’s also far better content. This post was slop.

This is all AI slop

How do these people keep up with technology? Those are basic skills now.

PuffPal, a platform that manages membership and age verification for cannabis retailers and clubs across Europe.

At least we’re keeping the children safe though by verifying ages. It’s worth giving up privacy for that…

Gift link: https://www.wsj.com/tech/the-taser-ceo-who-says-ai-is-the-fu...