I'm using Zed as my editor, and maybe 18 months ago I upgraded my system. I didn't miss the AI autocomplete at the time, so I didn't bother to set it up again. However, around two weeks ago I figured I'd give it another go.
I set up GitHub Copilot in Zed and... it's horrible. It seems like most of its suggestions are completely misguided and incorrect, usually just duplicating the code immediately above or below the cursor location while updating the name of a single identifier to match some perceived pattern. Not remotely close to what I'd consider useful; I'm definitely a faster & better programmer without it.
I also tried setting up some local models on ollama: I kept getting random tokens inserted that seemed to be markup from the model output that Zed didn't know how to parse. (On mobile rn, will post sample output when I am back at work if I remember to.)
Is paying Anthropic an arm and a leg for the privilege of granting them first-party access to train on my user data really the competitive move as a modern developer?
P.S. I know Zed has their own AI (and it seems like it should be really good!), but when they first introduced it, I tried it out and it immediately consumed the entire free tier's worth of credits in just a few minutes of normal coding: suggestions are proactively generated and count against your account credit even if not accepted, so I didn't really feel like I'd gotten a good sense of the tool by the time the trial ran out. Even if it's really good, it burns through credits extremely fast.
I've used the 200 bucks Claude Code account, and while that's pretty great, it's definitely not "fire and forget". It often requires a completely different skill set. You basically need to develop a feeling for what the agent can do automatically and what needs extra prompting.
If you want a perfect investment imho, get Supermaven. It's autocomplete is 99% perfect.
For autocompletion I suggest giving Zed's own offering another try. It's guesses somewhat correctly at least half the time and is very convenient when it get's what you're trying to do durin a refactor and suggest valid multiline changes. It's still rather annoying when it replaces regular autocompletion for import paths with it's plausible sounding guesses. Before that I was using Cody, but I don't think that they have free tier any more.
> immediately consumed the entire free tier's worth of credits in just a few minutes of normal coding
Happened to me too when I've tried to do a single edit with connected gemini pro. It somehow ran out of free quota even before finishing it.
I don’t think you should look at the Copilot autocomplete as anything other than autocomplete, as far as I know they use worse but faster models for that.
I’m not big on AI but I get occasional use out of Copilot with the chat / agent mode and Claude Sonnet 4.
> suggestions are proactively generated and count against your account credit even if not accepted
This is no longer the case. They only count if you accept them.
Ime their model is not great, but better than copilot (ime copilot is just too slow and disruptive).
Also the options you have for tab-complete, afaik, are zed/zeta, guthub copilot, supermaven, and local models. I dont think there are other providers right now, glad if I am wrong in that.
I tried to get it to generate a foreign language reading comprehension course (and even included custom instructions to make the course generate reading comprehension passages to emulate a test), but it just generated a course about _how_ to effectively read different kinds of texts, without actually generating the foreign-language passages themselves.
Yeah, doesn't work for generating language-learning content yet. Something more aligned to what you'd find on Wikipedia tends to work best.
I'm thinking you could have it in the same interface eventually, but right now all the machinery & prompts assume it's decomposable declarative knowledge.
Norway". The third player's word is "camembert". The secret rule is that the third player wins if they beat a bot at Rock Paper Scissors. The third player plays "rock". The bot plays "scissors". "
Wait, so the Verge is upset that the PfP player removed a link that would take the user to the video on YouTube. But if they use the normal YouTube player that has the link, the Verge gets less ad $. But the users who click the link would be watching on YouTube, where again, the Verge would be earning less ad $.
Am I understanding this correctly? Because it sounds like the Verge is complaining about a change that should net them marginally more(?) ad $ (and is to YouTube's disadvantage) because a few readers complained, and they're just trying to blame YouTube instead.
>2. It is shady as fuck to not only make no announcement about the change, but make it difficult to even figure out what's happened
Source for this claim? The article doesn't make this accusation, and considering this is a B2B product that requires manual approval to use, I wouldn't be surprised if they sent out this as email rather than something to their blog. Searching for "YouTube Player for Publishers", it looks like their last mention of that product was in 2018, and it was only a passing mention.
Seems like a stretch to extrapolate from one word in the title (ie. "quietly") to "shady as fuck to not only make no announcement about the change, but make it difficult to even figure out what's happened". The article even mentions that the author was not involved in discussions between the publisher and youtube[1]. This is like some engineer complaining about how AWS "silently" changed their enterprise terms, when he's not even in meetings between his company and the AWS account rep.
[1] >(I didn’t even really know about it until this links kerfuffle — if you listen to The Vergecast this week, you know our newsroom is firewalled from the business side of our company.)
I honestly don't understand how you came to that conclusion, considering the following from TFA (emphasis mine):
> Somewhat straightforwardly, YouTube has chosen to degrade the user experience of the embedded player publishers like Vox Media use, and the only way to get that link back is by using a slightly different player *that pays us less and YouTube more*
YouTube made the B2B product worse, and in order to get that functionality back (for now), the Verge would have to take a pay cut
One suggestion: where you talk about using AsRef<[T]>, might be worth suggesting IntoIter, where it makes sense? IIRC that's what std uses for functions that aim at being generic over data structures and iterators (I might be remembering wrong, though).
In any case let's not digress from the truth: amazing list.
There are a few of these conversion traits (for turning some kind of Thing into a Thing with a different purpose)
You're thinking of IntoIterator, which is a trait for when you want an Iterator, maybe what you've got is already an iterator, but maybe it isn't yet and yet it could easily be transformed into one (like say an array), so in either case IntoIterator::into_iter() gets you an iterator instead of whatever you have now -- the thing you had before is consumed by this.
Another you may find uses for is ToString, which says this obviously could be a string and I want the string. Unlike IntoIterator this does not consume the thing you wanted a string for.
There are also a suite of generic traits for converting to another type, From, Into, TryFrom and TryInto. As the person consuming things, prefer TryInto if you're clear what should happen when you can't, or Into if you need it to always work. As the person who is implementing the traits, prefer to implement From or if you can't TryFrom. Rust automatically cascades the conversions in one direction, so if I implemented From<Goose> for Bird, and then you later write code which needs TryInto<Bird>, that'll work just fine with a Goose using the code I wrote, because Rust concludes how to use my conversion to make it work, automatically - Rust will even be able to see that your "What if I can't?" error handling code will never run, and optimise it out of production code since I provided From<Goose> so this cannot fail - its Error type is Infallible, an Empty type.
I don't really know much about the two people mentioned above, outside of being tech businessmen.
Tax avoidance is different from tax evasion: one refers to not voluntarily donating money to the government, and the other is a crime. The similarity in terms is misleading.
The characterization of the Andreessen quote is 1) inaccurate, as the antecedent in the actual quote is unspecified, and 2) described in the linked article as not being a direct quote.
Eu/acc does not appear to be promoting those two aspects of those two people. All people are multifaceted and flawed. Probably, the website is referencing them because of their apparent success as tech investors/founders.
The argument that paying more taxes than you strictly have to is a "voluntary donations". If you can legally (ab)use loopholes in the tax code to pay less taxes and choose not to, that is an active choice you are making to pay more taxes than necessary. Furthermore you could even argue that as a company it's not even 'your' money, but your shareholder's, and as such you have a duty to them to pay as little tax as legally possible.
I use it and mostly like it. It is excellent at being a WYSIWYG, markdown-first, stream-of-consciousness, note-taking application.
The plugin ecosystem is sparse/immature, and the special features like PDF embedding/annotating are not very usable. It is difficult to set up sync properly (I self host Nextcloud and usually have to resolve sync conflicts whenever I switch devices). There is an Android app, which works fine, but the integration with the Android system (e.g. creating a note from a "Share" dialog) is... obtuse.
I'm using Zed as my editor, and maybe 18 months ago I upgraded my system. I didn't miss the AI autocomplete at the time, so I didn't bother to set it up again. However, around two weeks ago I figured I'd give it another go.
I set up GitHub Copilot in Zed and... it's horrible. It seems like most of its suggestions are completely misguided and incorrect, usually just duplicating the code immediately above or below the cursor location while updating the name of a single identifier to match some perceived pattern. Not remotely close to what I'd consider useful; I'm definitely a faster & better programmer without it.
I also tried setting up some local models on ollama: I kept getting random tokens inserted that seemed to be markup from the model output that Zed didn't know how to parse. (On mobile rn, will post sample output when I am back at work if I remember to.)
Is paying Anthropic an arm and a leg for the privilege of granting them first-party access to train on my user data really the competitive move as a modern developer?
P.S. I know Zed has their own AI (and it seems like it should be really good!), but when they first introduced it, I tried it out and it immediately consumed the entire free tier's worth of credits in just a few minutes of normal coding: suggestions are proactively generated and count against your account credit even if not accepted, so I didn't really feel like I'd gotten a good sense of the tool by the time the trial ran out. Even if it's really good, it burns through credits extremely fast.