presentation

no profile record

comments

19 дней назадdiscuss

I think also that anyone who's spoken Japanese for a while already has internalized that "si" === "shi" because there literally isn't the sound "si" in modern Japanese, as the other commenters mentioned it's often romanized both as "si" and "shi" in daily life, if you typed "si" into a keyboard it renders し, it goes on. The original comment on this thread includes one such person who literally didn't follow why "si" is wrong, and I felt the same way too as a long-time Japanese learner. It's a very "copy paste Western language concepts onto Japanese" way of conceptualizing the language, which is IMO a great way to set oneself up for great struggles when trying to learn a language that is structurally different, because it's not the right mental model.

23 дня назадdiscuss

So does that imply that there is a Goldilocks zone of prosperity where society are wealthy enough to afford good infrastructure but not too wealthy so as to become unable to build infrastructure cost effectively? Interesting that too much prosperity would actually lower living standards.

23 дня назадdiscuss

Curious, what do you dislike about him? I find many of his takes to be pretty reasonable, though he gets temperamental and snarky sometimes.

24 дня назадdiscuss

I primarily use Cursor, compared against Claude and haven't used Codex before - to me the benefits:

- Composer 2.5 is cheap, fast, and very effective; I don't even use other models that much anymore, as it's usually marginally better for way more cost, though sometimes I do for specific things like making better translations for our app than a coding-specific model could normally output

- It makes setting up and maintaining Cloud Agents super easy, the agent can basically set up itself and if anything changes that makes it not update properly, then it tells you and can fix itself easily

- Easy to move agents from remotely in the cloud to local and vice versa

- Easy to work with agents via either comments on GitHub, or via the web app on my phone (though it is relatively constrained relative to the actual desktop UI, which is a bummer)

- Code reviews with Bugbot is surprisingly good now vs when it first came out, as is the Security Agent, while being an order of magnitude cheaper than stuff like Claude reviews

- Automations are easy to configure and manage - crons, in response to repo events, etc. For example I don't use Renovate or Dependabot much anymore since LLMs can update deps and investigate subtle breaking changes much better than a dumb version bump script can

- Limits are obvious rather than Anthropic's mysterious quota amount that they don't explain at all

- Queueing up messages rather than the agent taking in new messages mid-work and then trying to mesh them together somehow - I find queueing much more predictable and easier to work with

- Plan mode is good too, but not particularly different than any other agent

- Easy to jump into the Editor view and actually go in and manually code things or interactively code with the LLM when you need to, since sometimes LLMs just suck at doing certain things autonomously. I'm not in the "stop coding bro" camp, I still like to take the wheel fairly often.

в прошлом месяцеdiscuss

I think it’s been banging it’s head against the wall for many months now lol

в прошлом месяцеdiscuss

I ran into this earlier, the commit history is the one maintainer and Claude pumping out like hundreds of commits 24 hours a day. Aside from the token cost, I’m wondering how this will turn out relative to the official tsgo native push.

2 месяца назадdiscuss

What about both? Artists want money, fans want entry, reserve a portion for hardcore fans and the remainder by auction. Artists get to sell their $10k seats to the rich while looking like they’re giving an amazing discount to their fans.

2 месяца назадdiscuss

Tried that, it just seemed way dumber this way unfortunately. And the zed UI provided 0 visibility whenever it was doing tool calls, and for some reason it kept running sleep 30 calls because it couldn’t figure out how to see the results of its own tool calls for some reason.

2 месяца назадdiscuss

If anything it’s a tool for making people outside of Singapore like/want to do business in Singapore, so if that makes it some twisted kind of utility then I guess anything can be a utility. Not like they have domestic flights.

2 месяца назадdiscuss

Yeah we just use Zod’s branded type and that pretty much handles it. No casts, use a refinement then slap a brand on it.

2 месяца назадdiscuss

That's the point, you need to make making money = improving society. If you think that's impossible in the USA, then the USA is doomed and maybe you should look for the door. I am not that pessimistic.

2 месяца назадdiscuss

Looks much better than when I last tried it! But I couldn't get it to work as well as Cursor for AI development, maybe I just need to get more used to it?

- I tried to use the Cursor Agent via ACP, it worked but it seemed markedly stupider than when I would use Cursor directly (saying that i18n strings were being used when they weren't, editing code differently than what I asked for, also when it is running terminal commands it seems to just say "Run Command: Terminal" and has no information on what's going on). Maybe I just need to not use Cursor Agent, but my company pays for it already so that's what I tried.

- Providing context is also cludgier - In cursor I often highlight specific chunks of code and add it to the AI context via Cmd+L, but I couldn't figure out any specific way of doing that with a keyboard in Zed besides clicking a tiny + button to add "Selection" to context, which got old fast.

- Maybe I just need to get used to it but reviewing code with the git integration is just hard for me to follow; one giant editor with every change in it is just harder for me to grok than showing each file one by one; so it was tiring to review the big changes produced by LLMs. Also, when you stage changes the file just stays where it is with a barely noticeable check in the checkbox in the sidebar - I prefer the behavior of Cursor where it actually moves the staged files to a separate section, but some kind of more obvious visual indication besides that perhaps would help. I did like the tree view, though!

- The tsgo and oxlint LSP servers kept crashing, which was frustrating. GraphQL LSP server also couldn't understand graphql.config.mjs, which is strange as that's supported out of box by graphql-config and works fine in Cursor/VSCode.

- I tried using a few of the different Edit Suggestion LLM options, but unfortunately Cursor is just way too good compared to any of them (slow, and just not very helpful in comparison).

- Just in general figuring out how to configure them is confusing, there's like 3-4 different places to configure agents and LLMs for different purposes, I found it very fragmented and confusing and the docs didn't make it particularly easy to set things up.

That all said, the performance was muuuuch better than Cursor. But the UX issues and general bugginess of ACP and these LSP plugins were impeding my workflow too much for me to tolerate it so back to Cursor for me. If anyone has tips on how I might make some of these better that would be cool, but if it's just inherent limitations then maybe I'll try again in 6 months or something.

3 месяца назадdiscuss

Did you read what I said? The whole Japanese system is for profit and the one of the biggest reasons for Japan’s system being so pleasant is that it is done for commercial purposes.

If the incentives are right American companies can make good things, but usually they are not so because of poor policy.

3 месяца назадdiscuss

Works in progress also had a great article recently (also discussed on hacker news) about how Japanese railways are private, profit earning real estate development corporations. [1]

Unfortunately, people from western countries have very negative views toward the privatization of mass transit despite the wild success that Japan has experienced. The model makes so much sense: if trains are just a way to get people to the real estate that you developed, then you’re going to make sure that the trains AND the destinations are really nice, which also turns out to be very lucrative (at least in densely populated areas) as a cherry on top.

And even worse, like this commenter above alludes to, it is trendy in the West to believe that real estate developers are evil, and that corporations that make money are sucking the life out of society. This kind of degrowth populism pretty much guarantees that the successful Japanese model is out of reach for most countries, because it is exactly the pursuit of profit that makes Japan’s system so nice - not some edicts from a benevolent and extremely capable government.

[1] https://news.ycombinator.com/item?id=47762060

3 месяца назадdiscuss

One example is that I have a fancy visualization in my app that is rendered in the server via RSC and just some interactive tidbits get sent to the client. If I packaged the whole visualization library it would have bloated my bundle size but instead I ship barely any JS and still get a nice interactive vector data viz experience. And the code just looks like normal react component nesting more or less.

3 месяца назадdiscuss

Well, to be honest, the results in Japan and China, where that isn’t the case, have turned out to be much better.

3 месяца назадdiscuss

China is giant and sprawling and they are able to do it.

That said this reply doesn’t actually address much of what the article talks about, most interestingly how rail companies are private and are also real estate developers. That thought process ought to make sense to Texans or something.

4 месяца назадdiscuss

Hating Palantir without having any idea of what they are is the trendy thing to do. Their leaders are toxic which doesn’t help the case, but the core issue really is just that in this political climate, people all over the western world don’t trust their governments, and it’s also trendy to distrust anyone making money, as well as tech companies - especially those involved in data and AI related businesses - so the fact that Palantir makes these distrusted actors more competent while making money doing it, is seen as siding with the devil.

So it’s a trust problem, if the government were seen as effective and worthy then I want them to be effective, which includes using the data they collect effectively. In this climate trendy people would prefer that their corrupt government is also fully incompetent to limit the effect of the corruption.

4 месяца назадdiscuss

I for one am not shedding tears about being able to waste less time with SOC 2 drudgery that doesn’t even make organizations secure.

4 месяца назадdiscuss

I broadly agree with this, it really is all about trust. Just, as a company scales it’s hard to make sure that everybody in the team remains trustworthy – it isn’t just about personality and culture, it’s also about people actually having the skill, motivation, and track record of doing good work efficiently. Maybe AI‘s greatest value will be to allow teams to stay small, which reduces the difficulty of maintaining trust.