Meanwhile, as a generalist who has a basic understanding of general things, everything from how to design efficient network protocols, to how cache lines affect the performance of sorting algorithms, without being a real expert in any of those things, I act as a constant course correction for AI agents doing work on my behalf, in a way that LLM context windows simply cannot replicate.

To give a concrete example, I recently used agents to build a specialized sync protocol that broadly resembles Dropbox. It's nowhere near as efficient in terms of how blocks are synced (because it entirely happens on a LAN and the cost difference is minimal), but I constantly had to make objectively more valuable course corrections on how the sync actually traversed the participating nodes. If I'd just let the LLM drive, it would have come up with a reasonably efficient algorithm (better than I probably would have done on my first try in the same timeframe) that would have had an obvious (to me) single bottleneck.

1. It's much more verbose about how it perceives the current state of things, i.e. "this is a large, well-documented project"

2. It's much more willing to trust its own judgement, e.g. fewer prompts to approve decisions

3. In terms of how long it takes to solve isolated problems, and the quality of solutions it proposes, it isn't meaningfully different from 4.7

YMMV, and maybe my view will change as I work with it more, but it feels like system prompt tweaks more than a real step forward

This is the distilled essence of a “first world problem.”

It’s obviously bad to eat super salty “ultraprocessed” food all the time, but it’s not like the salt is the primary problem

To take OP’s example, I’d much rather kids eat generously salted broccoli that is “optimized for taste” rather than unsalted mac & cheese, regardless of whether they just throw it away (which I probably would, too)

With a mass market electric sharpener and a reasonable knife I spend maybe 15 minutes/yr on sharpening and the knife + sharpener costs less than half this product

The marketing video seems to try to head people like me off, but it also seems to wildly overstate the level of commitment required to have sharp knives

(I do think the tech is cool tho. I just wouldn’t pay $400 for an 8 inch chef’s knife no matter how good it is)

If you have a bad machine get a good machine, but you’re not going to get a significant uplift going from a good machine that’s a few years old to the latest shiny

It wasn’t bad, and I’m sure I’d just get used to it if I picked one and lived with it, the same way I’ve gotten used to Apple’s dumb photo app

Using them side by side made it really obvious tho

Based on his progress posts it seems that ATproto is intentionally moving in the right direction and BlackSky has progressed to the point he’s asking for volunteers to move off Bluesky and try out his implementations

"Undoubtedly" seems like a level of confidence that is unjustified. Like Travis Kalanick thinking AI is just about to help him discover new physics, this seems to suggest that AI will go from being able to do (at best) what we can already do if we were simply more diligent at our tasks to being something genuinely more than "just" us

Old sailboats can be had for practically (and in many cases actually) nothing. If you’re reasonably handy and willing to learn you can do all the maintenance they require yourself

Boats can be some of the cheapest housing there is, even more so if you want to live somewhere picturesque

(There are, of course, significant downsides)