It’s not just engineers. Society has collapsing birthrates and huge deficits. Basically, we are demanding massive technological gains enough to bump GDP by at least 5% more per year.

I strongly dislike the “updating of versions” whenever possible. Versions are rarely better in all ways, makes things harder. Just make it version 2.6.

my thoughts on the email design: - Comparison is strange. One email has an image, the other text. Not the same email. - Hiding the previous parts of the thread seem good by default, but how do you easily get them back? - Where is "from" in new design? - Where is "to" in new design? - I do like expanding attachment a bit so you don't have to click twice to attach a photo (for example), but I'm not sure how often some of those options are used, may be too much. I could see a photo icon and general attach icon both showing. - Back arrow looks broken in new design.

> Not the same email.

I'm not even sure they're both emails. The first looks like a fairly conventional mobile email app; the second looks like a messaging app.

Not only does it not have a 'from' and 'to' field, it also doesn't have a 'subject' field.

That is great. Always nice to see sane IP decisions.

I’ve become increasingly frustrated with AWS because of this. They used to have a culture of providing constant price performance improvements. Not anymore. Every release has questionable improvements (for example, switching from r6 to r7 family of instances is more expensive with theoretically better performance but you probably can’t actually switch and save money). S3 costs haven’t gone done in a long time despite plummeting storage costs.

Very excited about the work done by 37Signals to encourage moving off.

In the 2037 timeframe, modeling trends doesn’t matter as much as looking at the actual players. I think odds are good because you have at least 4 very well funded groups shooting to have something before 2035: commercial groups including CFS, Helios, TAE, also the efforts by ITER. Maybe more. Each with generally independent approaches. I think scientific viability will be proven by 2035, but getting economic viability could take much longer.

Interesting study, good work by authors, I read it as little to no effect though. Confidence interval includes 0.14 kg of difference in weight loss. But I couldn’t read the full paper to show how they tested compliance between the groups.

For me, fasting (skipping breakfast or breakfast and lunch) and eating more vegetables (to feel full) are the easiest ways to hit calorie targets, but I suspect that is the only effect.

So the practical takeaway of this, and all diets, is focusing on what is the most sustainable way to hit calorie targets.

Without commenting on the cuts themselves, this article suggests its regulatory reform that is needed to keep up.

Yeah, 2023 I would expect no effect. 2024 I think generally not, wasn’t good or deployed enough. I think 2025 might be the first signs, it I still think there is a lot of plumbing and working with these things. 2026 though I expect to show an effect.

That's now what the AI boosters and shills were saying in 2021. Might be worth a refresher if you've got the time, but nearly every timeline that's been floated by any of the leadership in any of the OG LLM makers has been as hallucinated as the worst answers coming from their bots.

2024 was already madness for translators and graphic artists, according to my personal anecdata.

Anyone know of a way to just have a remote mcp server on top of a rest API? I want to do something simple to see if my users will care, but can’t find anything trivial (have public API, just want to try and wrap it in mcp for low cost). Tried stainless but it didn’t really do that (seemed to just generate boilerplate for me to install).

https://github.com/dcolley/swagger-mcp perhaps.