feifan

Feedback helps, right. But if you've got a problem where a simple, contained feedback loop isn't that easy to build, the only source of feedback is yourself. And that's when you are exposed to the stupidity of current AI models.

That's exactly the point. Modern coding agents aren't smart software engineers per se; they're very very good goal-seekers whose unit of work is code. They need automatable feedback loops.

Cultural reflex probably; lentils and tofu are displeasurable to most Americans

this assumes the application is hosted as SaaS, but if the application makes sense as a personal/"desktop" app, that likely wouldn't matter.

Thanks Mahmoud! Does the code review aspect of this demo resonate more with you than the authoring part?

Thanks! Getting all the possibly-relevant related code isn't too hard; the hard part is filtering down to what's relevant. So I think specific use cases (e.g. "how do I get from a known entry point to some method/property/etc") is gonna be key.

The company has stated that if we do not prepay the taxes by March 15, 2025, the RSUs will be permanently forfeited

Is it clear whether you'd forfeit your entire grant, or just some subset of shares that would correspond to a tax/withholding percentage?

Positioning/messaging matters a lot!

I'd suspect some sort of economic factor, or sampling … e.g. just people with lower specs submitting and bringing down the average

I read that sentence as saying someone came up with a cost estimate of $142.7b to clean up everything, and then there's a comparison to 3M's $53b market cap for scale/comparison.

Apps like Bear, Reeder, Anybox, Things 3, make switching devices seamless

I think Apple's whole point is that they provide the APIs so that apps can do this themselves. Which, yeah, is a bit disappointing; I'd love to have a 100% fidelity, flick-to-another-device type of Continuity built into the OS, but the APIs are at least a starting point.

Consider storing the data on Backblaze B2 ($0.005/GB/month) and serving content via Cloudflare (egress from B2 to Cloudflare is free through their Bandwidth Alliance).

(No affiliation with either; just a happy customer for a tiny personal project)

I'm working on an app to browse a codebase as a list of constructs — classes, methods, etc. It matches how I think about code: as a database of named entities, in which the implementation is often an afterthought — unlike text editors, in which constructs of various types, along with their implementation, all kind of look the same. It's still a very early project, but here's a short recording of some of what I've built so far[^0]. It's been a great way to learn SwiftUI, the Language Server Protocol, and FSEvents as well!

[^0]: https://twitter.com/tanagram_/status/1652164856563507200

I've seen data showing ~70% recovery on a Model 3, but I also can't find the source :/

The part that Siri is bad at will be commoditized — someone will open-source a GPT-4-level language model. And Apple's moat will be being able to run that on-device with all the attendant benefits (privacy, zero marginal cost to the company, availability in more scenarios, etc)

A well-adjusted person should be able to signal some degree of confidence about what they're saying ("I know that…", "I think that…", "Maaaaaaybe…", etc). That's an anspect that ChatGPT doesn't have.

Is your photo library on a spinning hard drive or USB drive? I had this problem before which the setup, and it went away when I moved my library to my internal SSD.

I understand and partly agree with your sentiment, but "offsetting" technologies/components are all over the place. E.g. combustion engines also need mufflers, catalytic converters, etc.

Cost-per-mile is cheaper from electricity than from diesel, and it's expected to have lower maintenance costs too (fewer moving parts, no oil changes, etc)

The displays have downward-facing cameras covering what looks like the blind spots next to the cab

It depends on your time horizon. Lots of people deride Chamath's SPACs, but several of them traded well above their initial $10 price for 1.5+ years until the market started dropping at the beginning of this year.

I do feel a warm, fuzzy sense of satisfaction when I get to light up all the cores on my workstation … but that might just be from the heat of the Xeons ;)

Both are true — if a team can figure out how the reduce the need for human intervention, they generally will. In the limit, what's left is the really-hard-to-anticipate/automate stuff.

One example that I use often — the Photos app on Mac has a built-in editing interface, but also supports extensions that allow third-party apps (e.g. Affinity Photo and Pixelmator) to take over that whole UI and present their editing UI inside the Photos app.

You'd probably want to look for small companies making something that impresses you (which may or may not be public). Hopefully they stay small for a while and preserve that culture.

Identify which of the files should be modified first.

I'm curious how many software engineers work this way — i.e. you need to make a big change, and you can determine a sequential order in which to make changes?

My working style is mostly jumping back-and-forth across all the relevant files and tests as I go, realizing new things that I need to change, and iterating until it all works end-to-end.

Am I the outlier, or is Maiao's expected workflow the outlier?

The hard thing about plain-text is that you quickly run into domains/problems that are too complicated to easily represent as plain-text. When that happens, you run into problems: the text representation becomes really verbose (e.g. XML) or exhaustingly terse/unergonomic (e.g. LaTeX, IMO), and you can only build anemic tooling.

I'd love to see the opposite — bigger keyboards, with more keys dedicated to domain-specific functions (something like this[0]). I think that would mean less cognitive overhead in general as you do a particular job.

[^0]: https://twitter.com/mwichary/status/1537631208414191616?s=21...

Back when I had two monitors, I got the same cables for both and I indiscriminately plugged them into my computer (I had a laptop, so plugging/unplugging happened ~daily). So the mapping between computer port and monitor definitely wasn't going to stay the same.

$30B is still A LOT in absolute terms

TIL nitrogen contributes to fertilization