> They are pointedly ignoring FPGA and ASIC. With the current model quality, would it really be so bad to burn Claude irrevocably on a chip and have a non-modifiable, cheap to mass produce, order of magnitude faster Claude-in-a-chip? This is what happened to Bitcoin ultimately, there are huge performance gains we know exist lying on the table just because they don’t exist for training. And even for training, TPUs make the first steps in that direction.
IT outsourcing was arguably a sort of precursor to vibecoding, and consulting companies are arguably still eating well cleaning up the legendary mess it created.
Yeah it helps make awful code decent, and some algorithms are better than others, but in terms of high performance code, locality, vectorization, and branching often matter much more big O.
More often the past mistake is rewriting software in a newer language.
Like I've worked places where the business ran off a layer cake of '80s tech, '90s tech that partially replaced the '80s tech, and '00s tech that partially replaced the '90s tech, and were now on their way to launch a big project to replace all that with '10s tech, a project doomed to run out of steam half way through (because legacy code got hands), inevitably leading to a codebase that consists of three failed attempts to rewrite the '80s codebase, and the '80s codebase.
As the functionality of the code was business critical, and no shift in behavior could be tolerated, they're never getting out of this mess, and would have been better off staying on '80s tech.
Honestly where do you even go if you want to get out from under this? The US was the option, but is clearly circling the drain. The EU is democracy theater at best, a democratic mandate that can be set aside any time it's inconvenient for Ashton Kutcher, and speedrunning the rebuilding of a new Soviet Union. Feels like a matter of time until they start building a new wall to keep you from leaving.
I don't understand how this stuff prevents me from MITM:ing the attestation and routing requests to another device that is clean and attested? I.e. how do you know that the remote TPM belongs to the device you're talking to?
The way we should define code quality is arguably how easy it is to affect correct changes to the code, that's hard to quantify, but ultimately the thing any code quality metric is trying to capture.
Based on that, it should be a pretty unsurprising conclusion as long as the code quality metrics you are using are reasonable; as long as the quality metric is good (within the context of coding agents), then this is the result we'd expect to see.
Well yeah, it kinda feels like you're using NFS, and the server you're connecting to is in orbit around mars and is using a pringles cantenna to get its wifi signal back to earth.
Like the vanilla file explorer experience is way worse than anything I've come to expect with stuff like CIFS and SMB.
I really don't understand how you can even create software that feels as bad to use as Windows Explorer. It's like it's barely attached to reality. There's this weird floaty delay in everything. You copy a file, or did you? You're not sure. It hasn't updated yet. Oh, now the copy dialog appears with this progress bar that isn't showing progress. The dialog just sits there. Is something happening? I don't know. Many seconds later the dialog closes. But it hasn't showed up in the window yet... oh, now it did!
How is that even possible, especially with modern hardware? Like you'd almost have to build the file explorer around like a sqlite-based message queue with a 1500ms poll interval to get performance characteristics like this. Absolutely insane feats of architecture astronautism are no doubt required for this to happen.
It's a thing in some Germanic languages. Instinct is to merge nouns into word, e.g. 'lawnchair', but that gives you a red squiggly line, but 'lawn chair' also looks wrong, so 'lawn-chair' is the middle ground.
I'm funnily enough 40. You can definitely overdo exercise, but going from zero to some is generally very good for energy in my experience. It generally lets you turn that sort of tiredness you can't rest yourself out of into the sort of tiredness you can.
For energy, it both requires and pays dividends. It's a bit like working out in that sense.
I think my intended takeaway was that you really don't need to have make the thing you're studying take a lot of time, that daily consistency matters more than pouring hours into practice and obsessing about it.
Though in general, I do still think it's the phones and media diet that is the problem with the sense of lacking time.
Few years ago I had a full time job I felt like I had no time. Then I had a part time job, and I still had no time. Now I'm self employed, with nobody to answer to, and I still often feel like I have no time. Like damn, to get more time than I actually already have I'd need to move in next door to a black hole. Though when I unplug, then holy crap do I suddenly end up with a lot of time.
https://www.marginalia.nu/ <-- main website
https://marginalia-search.com/ <-- search engine
[email protected] <-- email