All projects done with Typescript, and the same tooling. The creativity of the LLM is quite biased. I would expect more reasoning and choosing other languages, platforms, libraries, etc.
> I wonder whether we’ll see a secondary effect in the resale market.
I'm paying more on ebay for thinkcentre tiny and thinkpads - 12th gen intel and newer.
Refurbished spinny drives have been steadily climbing - up 50% since late last year. That's on top of the 20% mystery jump that happened in the last week of 2024.
We already are. Check eBay at the component level, which is showing it quite clearly. Look for secondary/reclaimed/refurbished components to backfill the gaps too.
Also be aware that this stuff whipsaws, if OpenAI actually takes posession of that memory and decides they can't use it and dumps, we're going to see a crash. Likewise if they back out of the deals with the memory fabs (or fail and default). There's some scary volatility on the horizon.
One of my trend-following and easily-influenced sisters was quite locked into the Apple ecosystem (iPhones, MacBooks) with her young family, but just over Christmas last year, I spotted them with a refurbished ThinkPad (T490 or thereabouts), with a plan to buy another refurbished ThinkPad.
I /hope/ to see the slowdown in the new phones market affecting the pumped-up chatbot market. To reference Agent Smith, what good is a chatbot if you can't speak?
I love _Heat_, EXCEPT for the ending. (Spoilers) It just feels too forced as "the good guy has to win". So, I always stop it when the good guy cop says "He's gone!". It's much more in character with the bad guy; basically his entire career up to that point was being conservative and playing the long time, and I just didn't see the motivation for him to go after the cop and reverse his core policies. IMHO, it's a much better movie that way.
The funny thing about Calibre, which I've been using since the beginning, is that the app always has updates—I think they release one almost every day.
It's a bias to think the problem is people not knowing how to architect a good CSS solution. The reality is projects are maintained by multiple people over the course of its life. OOCSS only brings technical debt and constraints, and not to mention bloated CSS files. A good explanation why OCSS is bad https://www.fcss.club/manifesto and why functional CSS or atomic css is better in the short and long-term https://www.fcss.club/why. Tailwind just brings more complexity.
Can someone give me a reason why paying an engineer a yearly salary would ensure that they contribute a truly valuable piece of software that the market will use? Why should money be given based on years of experience? Is experience really a good indicator of ability? I know many engineers with 10+ years of experience who are mediocre.
What happens if that well-paid engineer decides they no longer enjoy software development? What if they need more than a year to deliver results? What if they produce something useless for the market?
This Keynesian approach is full of unknowns—it looks great on paper, but we've already run these experiments in the past, and they simply didn’t work.