Fig (YC S20) – Adds IDE-style autocomplete to your existing terminal
fig.io
fig.io
https://news.ycombinator.com/item?id=31113396 (76 points/9 days ago/63 comments)
And their Launch HN less than a year ago:
https://news.ycombinator.com/item?id=27277819 (559 points/11 months ago/378 comments)
Surely this has organically risen to the front page, unaffected by it's ycombinator backing!
when they appeared here last time, i've written a negative comment about them
my comment was then pinned to the bottom of the page by super-mods, even-though it had the first spot with 30 upvotes
dang (HN moderator) has actually confirmed that my comment's ranking was manipulated, but declined to answer any further questions
just imagine how much of the HN comment section is manipulated by these elites
i guess some YC-investor isn't happy that us regulars are devaluing his investment and downranks negative comments so that people will give Fig a try after reading bunch of good ones
I don't think this is the proper use of the word. These are just people with vested interest in making a product more popular - it has nothing to do with any "elite."
"a select group that is superior in terms of ability or qualities to the rest of a group or society."
the people who downranked my comments, were indeed elites, because they're a select group of privileged users with the ability to downrank
millions visit HackerNews every day
when those people see Fig ad here, first thing they do is check what the people are saying about it in the comment section
now, imagine that instead of real feedback, you're mostly seeing cherry-picked comments praising the product
it certainly shapes your opinion of the product and makes you more likely to try
I don't know why you are 'a regular' if you have such a toxic opinion of and attitude towards HN.
Also although this is not going to sway you, I at least want to mention it as I found it interesting when I came across this - see the response and embedded search link in the FAQ question "Are negative stories about YC suppressed on HN?" here [1]
Edit: also want to note that I'm thankful for you and others who bring up the privacy issues because I likely would have looked at the fig.io page, thought 'hey this looks cool' and upvoted the post without thinking much further... which is likely where a lot of the upvotes come from. So IMO putting things more lightly (even in your original response 9 days ago) would have sufficed. Perhaps if you had qualified what you meant by fig relying on "free labour" (it just sounds like a hectic accusation) your comment would have stayed at the top. Though overall I agree it doesn't seem like the kind of comment that should have been suppressed and it's good to push back against that a bit.
Most lurkers and users think YC companies make it to the front page organically and based on merit alone.
Other than that, YC companies have to compete for attention on HN along with everyone else, and I can tell you that most find it maddeningly difficult.
In some ways we use moderation power to YC startups' disadvantage - we moderate criticisms of YC startups less, not more, than usual (https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...).
For example, the tedious angry-generic-repetitive sort of subthread that tends to get upvoted on most common topics is exactly the kind of thing we downweight as part of standard moderation, but we do that less when a YC startup is the topic. That's not because such discussion is any higher-quality in those cases (it isn't), but rather because people are quick to jump to the conclusion that we're doing nefarious things and we need to be able to answer those concerns in good conscience.
The fact that there's all this drama about one such subthread having gotten downweighted is actually an indication of how we don't do that. It happened, but only by mistake.
But if some users can down-weight comments and that happens on a thread about a YC company, I hope you can see why it would raise some concerns. I personally didn’t know some users had this ability.
It doesn't work perfectly though—the system is just too complex. In the case we're talking about, the user (not mod) who downweighted the comment was quite aware of both those rules and of their precedence. They just didn't know that Fig was a YC startup. There are thousands of YC startups and people can't remember them all.
Since mistakes are inevitable, I think all we can do is be open when they happen, fix them if we can, and explain our general principles are. That's what I've been trying to do in this case, and if there's a question I haven't answered yet, I'd be happy to. None of this is secret.
Actually the dream would be to have a community system that could regulate itself without any such special moderation, and the experiments I described upthread (https://news.ycombinator.com/item?id=31227642) are all about inching in that direction, but it's just a hard problem that we have no general solution for.
This is the first time I've seen something like this on HN though, so if anything it might show how good the HN policies normally are. I just wasn't aware that some members of the community are more equal than others; now that it is clear and communicated, I can make better decisions myself.
The intention is to try to inch in the direction of a more self-regulating community while improving the quality of the discussions. Improved discussion quality is by far the most important thing and as far as I can tell this experiment is a big success in that department, so the essence of the HN is definitely not being diluted by it. If it were, we'd stop it immediately.
Not sure if that helps or not!
Edit: if anyone wants to see a typical example, here's a top subthread that one of the users we're talking about correctly marked as offtopic: https://news.ycombinator.com/item?id=31231858. That's the kind of pointless flamewar stuff that tends to get stuck at the top of a page. Downweighting it makes everything better, and having more eyeballs out there to catch it sooner is super helpful. That's the intention behind this experiment.
For example, it used to be that only moderators could unkill a [dead] post and restore it to full visibility. Then we started an experiment in letting non-moderators do that, and this turned into the 'vouch' feature which allows most HN users above a certain karma threshold to have a say in which posts get unkilled. That was the biggest success to date - it has made a big difference both in thread quality (many interesting comments become visible) and community happiness (there's less resentment about posts being dead than there used to be).
That was a great outcome, and we're always thinking about more possibilities for such things. The current experiment is very much in the same spirit. Personally I would love to find ways to extend these things to everyone (most likely above a certain karma threshold) but it's not often obvious how to do that.
Another example is the second-chance pool (https://news.ycombinator.com/pool, explained at https://news.ycombinator.com/item?id=26998308), which gives submissions a random placement on HN's front page.
Being flagged / downvoted by users isn't "manipulated", that's just how a upvote / downvote based website works. The comment was likely downvoted because it was just plain out wrong as the first comment to your post already says.
> YC companies get tips and notifications when they're about to get featured
> We don't.
you're missing the point!
the comment wasn't flagged and it wasn't downvoted
in fact the opposite is true as it has 30 upvotes, upvoted by regular users like you
the comment was downranked, so it only appears lower on the page
this feature isn't available for regular users, only for an elite group
No, the point is that it was downvoted / downranked / flagged because it was a low quality comment that is spreading FUD and it doesn't matter who did it.
why isn't HN respecting user's opinions and manipulating the positions as they wish?
Because certain comments can be quite agreeable, but don't encourage interesting discussions. Your particular thread had more discussion about bitterness and etiquette than content. Even though your points are valid that's not the kind of discussion one wants to have at the top of any article.
Derailing the comment section about this article with your own agenda isn't particularly helpful either. But at least I appreciate that I now know about that feature.
but what happened though is that it wasn't legitimately downvoted, it was moved down the page
check my history, i didn't get 5000 points for posting "inflammatory, wrong, inciteful, off-topic, meta, or similar content" did i?
I checked the post in question at your request, and to be honest they way I read it is a bit off topic picking fights or meta discussion around the politics of HN. I read it as a bit inflammatory as well. Some of your responses here seem kind of similar, maybe you’re experiencing it differently but it seems like you are angrily responding to almost every branch of the threads here. But that’s all just my personal opinion of what I’m seeing.
Since HN's mandate is not to be that, we have to have countervailing mechanisms to compensate for that weakness. Otherwise the system would get stuck in one of its failure modes. The countervailing mechanisms include things like flags, software mechanisms (including automated up- and down-weights) and moderation (including manual downweights and other interventions). All this is happening all the time. If it weren't, HN would be a completely different place.
None of this is secret; it's not all transparent in the sense that we don't publish a full moderation log of all the factors, but we're happy to answer specific questions.
Just to be clear: I don't think there is anything outrageous about such a system. It's probably positive for the community.
my comment was on-topic and had valid points contributing to the discussion
the generic/offtopic is just a cover up
in fact why don't we just downrank your comment right here and mark it as generic or maybe off-topic? /s
We cut users a lot more slack when they're making shit up about HN, YC, YC startups, and so on, but the amount of slack we cut is nonetheless finite and it's time you stopped.
Lest we hear more about how you're being repressed: people criticize HN, YC, and YC startups all the time here with zero problems. You, however, have derailed more than one discussion now by inundating it with dozens of sensational, false, and/or offtopic comments. That is not what HN threads are for, and even though we moderate HN less when YC startups are the topic, we eventually have to draw a line.
Actually I have the impression that you're rather a good commenter in general and this isn't personal in any way, so if you want to fix this, it won't be hard.
Did you even read the comment?
> even-though it had the first spot with 30 upvotes
The comment had upvotes and a mod confirmed that the ranking was later manipulated, meaning it didn't lose ranking because of down votes.
My lord, it's one think when people just read article titles and skip the actual article, but it's incredibly immature and perplexing when people just read the first line of a comment!
Please don't jump to criticism and read the things properly, if not, try to refrain from commenting .
As Dang said in comment you linked: “That's no fun and does nothing for the community.”.
but
the reason i've got angry is because the original comment i made was moved to the very last position
not by downvoting, but by undemocratic rulers of HN
the angry comments in thread came after that accident happened
No, not everyone wants an ideological free-for-all.
If it's valid to criticise a critic, it must also be valid to criticise a critic critic.
The merits of a given argument or action do not rest on "because daddy said so". There is and must be a daddy, but no daddy is above critique themselves even though their job is to moderate others.
IIRC, the downweight was correct in principle, because the subthread wasn't about the particular story, it was just a generic complaint about the general topic. We want those downweighted because, regardless of topic, they have a strong tendency to sit at the top of a thread, accumulating upvotes and choking out more specific and more interesting discussion.
It was a mistake to downweight it, though—not because the subthread was good, but because the topic was a YC startup. By far the most important rule of HN moderation is that we intervene less, not more, when YC or a YC startup is the topic (https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...).
The user who did the downweighting didn't know that Fig was a YC startup, so it was an honest mistake. Certainly none of this has to do with trying to suppress criticism of YC or YC-funded companies. We know how to take our lumps and do so all the time. We also know that community good will is the only value that HN actually has, so doing anything to jeopardize that would be really dumb.
really? any examples you can share?
I'm all for keeping Rust discussions fun and curious.
sometimes you can get 2, 3, 5 times more upvotes than another post in half the time, if some HN admin/mod don't like your post then it's artificially ""downvoted"" and leave the front page(s)
he's right to be angry and hostile
Here on HN, we have something absent everywhere else: A moderator who treats you as a person and spends time and energy to explain why something is the way it is. You may not agree with the reasoning all the times but almost everywhere else you don't get the reasoning and only the results(shadow ban, censorship, lock out).
Can I offer you something? If you believe that HN is not run fairly, collect the evidence and write a thorough article about it and post it here or on Reddit and other places. If you articulate your ideas properly and if you have a point I bet you will receive traction and response from YC people.
HN is their marketing asset
otherwise how is it possible that a niche product of a YC company right now has the first spot with 400 upvotes?
I for one upvoted because this genuinely seemed like an interesting feature. It's clearly hitting a pain point for me as a developer, whether or not it actually ends up solving it satisfactorily.
> There are no down arrows on stories. They appear on comments after users reach a certain karma threshold, but never on direct replies.
The commenter here's upvote counted, my downvote did not count.
And I won't just use any other available form of attack like falsely accusing the post of some wrongdoing just because I want to criticise something and that's the only way to do it.
I did leave a critical comment but that is a seperate issue because you can leave a positive comment and upvote. So the ability to comment nulls out.
It's not a secret that YC startups get promoted and placed on the hackernews frontpage. If I remember correctly it looks different than this, so what exactly are you accusing them off? Using a hidden promotion mechanism instead of the known and accepted one? That makes no sense to me! If they have the ability to promote YC startups on the frontpage and it's generally known and not done in secret, why should they do it in secret here?
I think it legitimately got upvoted here. Maybe with some friends&family help. But not by some hidden moderator action.
I certainly found it interesting, but I would also never use a tool that privacy-violating.
why are so many people upvoting but only so little commenting? how is it that only a minority commented positively the product?
this isn’t like a typical #1 post
makes you suspicious doesn’t it?
We moderate HN less, not more, when YC or YC startups are the story*, but 'less' does not mean 'not at all'. When users turn threads into shitshows with false accusations and off-topic drama, we have to do something about that, YC or no.
* https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...
look at posting history of Fig, always the front page, always the hot spot
work for free for HN?
"The smart way to keep people passive and obedient is to strictly limit the spectrum of acceptable opinion, but allow very lively debate within that spectrum - even encourage the more critical and dissident views. That gives people the sense that there's free thinking going on, while all the time the presuppositions of the system are being reinforced by the limits put on the range of the debate."
The problem appear to be people have a high expectation from HN. People think that HN to be a discussion platform. But it is actually a PR platform implemented via a discussion forum.
Understand that and you ll have no trouble.
Also, I tried fig the first time round and was not impressed.
> Not sure if I can disclose further than that.
how hypocrite of you to confirm you know more, but don't want to give out any information
why are you confirming then?
You have a very curious definition of super-mod. One would assume a super-mod needs to be a mod first.
> confirm you know more, but don't want to give out any information
People are sometimes entrusted with non-public information they aren’t at liberty to disclose, at least without asking for explicit permission first. Still, when they see false accusations thrown around they sometimes would like to do something. I thought this is a forum for adults, that much should be obvious.
what matters is that there's a selected group of users, who can leverage capabilities unavailable and unknown to anybody else in the group
again, if you won't be disclosing any info, why do you mention that you have anything to disclose?
They've said enough, and there simply are situations where that's all you can say, and it's neither pointless nor wrong, nor even unusual or all that remarkable. It's common even.
maybe it’s just 5 people, who knows?
holy cow
The details would be even better, but I can think of at least a few ways a person might become aware of a fact like that without being able or morally obligated to disclose more than that heads-up, and without even themselves having necessarily ever done anything wrong.
It's not the DaVinci code.
> without the involvement of a “super-mod” or anyone affiliated with YC.
We moderate HN threads less, not more, when YC or a YC startup is the topic (https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...). but 'less' does not mean we don't moderate at all. Until now we've moderated your comments much less than we would if you had been behaving this way about any other topic, but enough is enough.
Standard HN moderation practice, when users do this kind of thing, is to penalize and rate-limit their account. I'm going to do that now. (Edit: I changed my mind later. See the explanation below.) Since the reaction this will educe is predictable, I might as well answer it now: no, this has nothing to do with criticizing YC or YC startups. It's because you've been posting a plethora of low-quality off-topic comments and otherwise breaking the site guidelines. This is standard HN moderation and the only thing different in your case is how long we held off doing it.
Actually, IIRC, you're rather a good commenter apart from this, so if you want to fix this, it shouldn't be hard.
Edit:
For some reason I realized later that I didn't feel good about penalizing your account. It's not that what I said about standard moderation practice was wrong or anything—I just have a feeling that you're a good HN user coming from a good place and that there's probably a much better way to fix this. So I'm going to go with that instinct, and have removed the rate-limit etc.
I would feel similarly concerned if Twitter or Facebook moderated conversations on their partners or competitors by suggesting that some users weren't intellectually curious enough, or that they had entrenched positions, or that they were full of negativity, or that they weren't "fun" enough. In this perspective, judging a user by the aforementioned values seems to be the wrong tier of values for the stakes at play. I feel that community trust is what is at stake here.
Also, you have mentioned before how HN moderation is not secret. Is my account penalized in any way? Given that some users have been hellbanned before, I do feel uncomfortable with the degree of perceived transparency here. As a regular HN reader, it is not my expectation that users shall receive transparency when they are penalized.
One of my takeaways from this conversation with the user <ushakov> is that they were initially opaquely penalize, and that we would not have received clarity on this issue had you chosen not to personally respond.
Re your specific account: your votes are downweighted. We do that when an account has upvoted too many stories or comments that break the site guidelines—particularly the sensational-indignant type of post, which is by far the biggest problem on HN and therefore the thing we try to dampen and downweight in every way we know how. If you want to commit to upvoting fewer such comments and/or stories, we'd be happy to take that penalty off your account. Other than that your account isn't penalized.
Re transparency in general: that's a complex question. We take the following approach: we don't publish a full moderation log of every decision; we do answer specific questions when people have them. I'm not saying that's a perfect solution (it isn't, and we have ideas for making it better), but this is a super hard problem to get right. It's popular to argue "everything should all be completely transparent all the time" but having been at this job for a decade now I'm extremely wary of the unintended consequences that would have. I'll try to dig up some old links about that in case you or anyone cares. Edit: ok, there's https://news.ycombinator.com/item?id=30141232 and a bunch more at https://news.ycombinator.com/item?id=27307075. If anyone reads those and has a question on this that I haven't already answered, I'd be happy to know what it is.
---
I've always been surprised when HN took moderation actions against my account, so it's hard for me to commit to not abusing when I don't have a clear mental compass about what is what. I just conduct myself in a way which aspires for personal authenticity in discussion.
I feel sad that the sum of my behavior on this forum is considered problematic and requires penalizing.
Re the penalty: I don't want you to be sad so I have removed it from your account. It's kind of hard to get into the details because we don't discuss specific vote data.
* needs email to download
* needs email to start
* spammy email subject with unicode fire in front (lost in HN): " You've been invited to Fig!"
* using trackers in email and app
* macOS only
Meanwhile, after reading comments here, I eventually added fzf completions to my shell and I gotta say: it's quite good!
No offence to the founders - but I am not going to trust access to my entire filesystem by a proprietary software just based off some projected goodwill + marketing.
Thank you.
screw you fig, thank you for enlightening me.
Fig completions are made by a group of unpaid volunteers and are incompatible with any other completion system
Fig pretends to be developing the product collectively in the open, but don’t give any code back to the community
of course they’d welcome everyone who wants to be an unpaid volunteer and improve their product in their own time without any expectation of a return
https://github.com/withfig/fig/issues/35
is it morally correct to exploit your contributors trust like that?
new slogan: "it's like Grammarly, but for code!"
meanwhile, just like Crypto AG, Grammarly (so useful btw to have a program that gives full access to any text-editing a person does, outside of the browser) is a small NSA project. but seriously, the onus is really on these companies to disprove this, and they have been found wanting.
yes it does sound like a conspiracy theory... but we know better now thanks to leaks involving Snowden, Crypto AG, etc.
We did nothing to make it rise. We do that for Launch HNs, a special case described in the FAQ (https://news.ycombinator.com/newsfaq.html), but otherwise YC startups are subject to the same community whims that everyone else is. We also don't do any special penalizing of negative posts about YC startups (just the opposite, in fact: https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...).
You guys need to understand that while you dislike some things about this product, other HN users find it interesting. That appears to be all that's going on here.
> Inevitably, there will be users who, for whatever reason, don't want their usage of Fig to be personally tracked. Of course. It's your data and software on your device. You shouldn't even need a reason. We want to cater to these users. And we will. But unfortunately, we are so early on in the process of building Fig that we need to be able to speak to our users. De-anonymising telemetry for the time being while minimisng the events tracked enables us to do this in a non intrusive way. But it is of course not for everyone. Therefore, once we reach a critical mass of usage in the coming months, we will then anonymise all telemetry, making Fig more accessible.
It has been several months since this was written. Are they still collecting de-anonymized user data?
However, during setup, you need to "sign-up" by sharing your email - you can't skip this step as far as I could tell.
they still send a daily ping
> Telemetry: Fig has basic telemetry in order to help us make product decisions. We currently give the you option to opt out of all non-essential telemetry by running `fig settings app.disableTelemetry true`. This removes everything except for one daily ping. We use this ping to help us understand how many people were using Fig.
If you assume he's contributing here in good faith (my assumption) then I'd say it's safe to say that it does in fact disable telemetry.
In Europe you must be able to prove that your software was designed from the ground up with enough safeguards for user privacy, as opposed to considering privacy requirements as a bolted-on afterthought.
I won't flag because I don't think it's wrong (although it approaches qualifying as spam, but, technically not quite, it's topical for this site, and I believe functions as advertised, and I'm sure a few mac users may legitimately find it useful), just bad and of no value and "do not recommend" and I'd merely like to do the opposite of upvoting, which is neither accusing of wrongdoing nor abstaining.
Considering it's a YC project and a commercial one, it's at least "spammy" if not quite 100% qualifying as pure spam. It's somehow, in a way I'm struggling to articulate, worse than merely not interesting to me.
It's so right on the edge. I sat with this comment open for a few minutes waffling between submit, or cancel and flag. It can be argued convincingly either way.
What's so bad about this post?
But this comment was about exactly what it says.
I'd leave some other comment about Fig itself but it's mostly all been said well enough already and I see no value in being redundant.
For the record, a reply where I added a few points https://news.ycombinator.com/item?id=31223471
I was then presented with a take home which was to make a minimal clone of https://explainshell.com/ using the Fig API. Needless to say this task was not only overly time consuming but it was extremely disrespectful to receive a reply over a month (!) after I submitted the task. They cited that my submission was “lost” while moving to a new system. If their management of new hires is this poor I shudder to think of what their internal management is like. Awesome product can’t say the same about the team though.
The same thing with wrap[0], to have autocomplete, somehow giving special permissions in the OS is not enough to function; I also need to give my identity and communication details. It feels wrong to me.
Edit: A nice alternative based on ZSH's native featureset: https://github.com/marlonrichert/zsh-autocomplete
IMO that's totally OK. I'm glad people are building stuff like this and if it requires an account to sign up, I am happy to do that.
> why I'm not allowed to use it without signing up?
I want to politely raise this question – is it not rude to have expectations for something you didn't build and be entitled that it to be offered to you for free? You're more than welcome to ignore it and find dozens of terminal emulators out there.
Well, Fig or any other free to use software wasn't commissioned upon my request therefore I don't think that I owe anyone anything. Engagement is worth quite a lot thise days, they earned my engagement by offering me something but did not disclose the requirements to use their product upfront so I ended up not benefiting from the product.
Think of this, you are walking down the street and someone really nice offers you a free meal. You accept, sit on the table and the food arrives but to let you eat the food they ask you for your contact details and a copy of you key for your home. You say, no thanks, would rather pay and they say no, it's only free meal in this restaurant.
I'm not sure I will accept it and assume that they will upsell me a boiler.
Okay, there's probably nothing nefarious but they stole my time and attention without giving me anything back. The restaurant shouldn't have advertised free meals but meals for key and contact details. I would have had gladly skip the offer, so they can concentrate on those who really need the meal and are willing to fulfil the request of the restaurant.
The correct analogy would be if you ate the meal for free but didn't want to provide your contact information so they can upsell their catering services to your business in the future.
All they want is a contact information to sell you stuff in the future for not paying today. You can use a anonymous email address and hide your identity. But, since this is a Terminal emulator, there is always a risk of privacy.
This is all pretty obvious. The argument really is "I want to have this thing for free, but I don't want to give anything back including contact details to further contact me".
Well, then how do you suggest a good business plan for building a commercial product if not having some ability to contact your users?
The contact details requirements are mentioned for the first time after install and just before start using the software, this creates a pressure to give my details. Probably that's the idea anyway(a dark pattern). This makes me trust them even less.
We are in the age of high profit software, be it through sales of product(x$ to use y), sales of extra features (x is free but this specific feature is y$), sales of services(x is free and your life can get easier if you use it with y for only z$/month), land grab(x is free for you because it's paid by y, who hopes to make money later somehow), data grab(x is free for you because we use the data collected here to build y), sales of access to people(x is free but we let other people reach you) or straight up criminals(you pay us x$ or y happens to you).
We can no longer talk about thankless customers of free products. There's even no longer free text on the internet, people write stuff only to show ads or influence your decisions(political or consumerist). The customer must understand how they are paying for it.
To give someone the benefit of doubt, they need an explanation for it. Clearly, contact details are irrelevant for the apps' primary function. They can ask for a license after a trail, I guess(free license obtained by signing up). They can let you use the app and offer you extra functionality(like syncing history between devices) through an online account.
It is not O.K. to make people install your app and only then require this one thing, irrelevant to apps functionality, to let you use the app. That's a known dark pattern.
How will Fig make money?
We get asked this question a lot and so as a final note to our launch, I thought it would be good to address it. First, Fig will always be free for individuals. We want users to feel confident that even if they sign up to Fig and become the biggest power user, they will never be charged. Our pricing model is inspired by GitHub's: For individuals and open-source projects, Fig is completely free. For team products that require hosting, Fig will charge a small fee per person per month.
Of course, not all projects succeed - most fail. Should this happen to us and we fail to gather the number of paid customers necessary to make the product self-sustaining, we will try to sell it. If we succeed, everything we said before becomes null.
Make of it what you will.
they offer a product for teams
teams can create private completions for Fig
in order to use the private completions, your whole team have to be subscribed to Fig
basically, their business model is vendor lock-in, because the completions you create for Fig don't work outside of Fig
But also autocomplete often suggests the wrong thing but I’m typing too quickly to notice, and then I hit enter and get the wrong thing filled in.
What I want is just non-automatic completion - which we already have. Ctrl+R gives me exactly what I need when trying to find some odd command I ran last week. Add in fzf integration and it’s perfect. Tab completion covers everything else. When I don’t want completion, it stays out of the way. In an IDE, you can usually bind completion suggestions to Ctrl+Space or something and then turn off automatic suggestions, or something like that.
Maybe you can set up this tool that way too - then it might be nice to try out. But as I mentioned generally I find the terminal to be pretty ergonomic already for completion stuff.
Granted, you shouldn't need to fiddle with it to make it usable, but i really didn't mind spending some time to get to know better the tools i'm using every day.
For the intellisense/suggestion options you can search for `editor.suggest`.
If that's not enough, you can arbitrarily modify the UI of the editor using this extension https://github.com/be5invis/vscode-custom-css
My config file is like 1500 LOC long and it's a bit messy, but in a quick search i found i severely limited the available options, changed the position of the box so it does not cover the text immediately underneath, changed some colors and put some suggestions as inline hints instead, among other stuff:
"editor.snippetSuggestions": "inline" "editor.suggestSelection": "first" "editorSuggestWidget.background": "#00000088" "editorSuggestWidget.foreground": "#ffffffff" "editorSuggestWidget.selectedBackground": "#9960FF33"
I've done battle with the VSCode autocomplete many times over the years, trying to stop things from popping up automatically when I'm typing. My escape key has gotten more of a worked out over the years than a vim user on speed. I've twiddled every option in Settings but it has defeated me every time.
Yesterday was my latest battle and I think I have the upper-hand. I set the "millisecond delay until show" to some very large value and it seems to have worked. I can now typed without stuff flashing and popping up all the time.
I came back late yesterday to the computer. VSCode was still open, and showing one lonely autocomplete pop up on screen. pfff.. Time to add another zero to the delay setting I guess...
Eclipse, Netbeans, Visual Studio, Monodevelop, VSCode, would be the ones I can recall from the top of my head.
edit: oh, it's a private company and it's mac only. Something about proprietary extensions to my command line is a huge turn off. Functionality itself looks great though.
I think almost straight copying landing page designs and layouts is basically fine and common practice. Using the same or very similar copy... not quite.
It'd be great if we had a template web site format so we know where everything is!
https://www.freecodecamp.org/news/fzf-a-command-line-fuzzy-f...
That said, I feel the need to support anyone trying to improve terminals.
But obviously, you can't turn that into a profit machine.
If someone could pull the autocomplete list out of fish and style it as a floating window that moves with the courser, that would be wonderful. Sounds more like a tool developed by one person than a venture funded startup though.
I wish I could try both Fig and Warp, but I can't: giving my email to a terminal is insane and I would never do that.
I wonder what is it that two similar projects popped up recently for quite non-business lucrative usecase but both require something that the very target audience will mostly never agree to?
(If you don't like bash's simple autocomplete, try other shells like zsh or fish.)
It is useful. But I really don't like waiting for cursor to appear when I want to type the next command.
Haven't found anything better than Ctrl+R.
Fig can be used with your existing terminals like iTerm, terminal or VSCode
Warp's landing page:
- Product Hunt #1 Product of the Day
- The terminal for the 21st century
- Warp is a blazingly fast, rust-based terminal reimagined from the ground up to work like a modern app.
- All cloud features are opt-in. Data is encrypted at rest.
- Fully native, Rust-based terminal. No Electron or web-tech.
Fig's landing page:
- Fig adds IDE-style autocomplete to your existing terminal.
- Works with your existing terminal & shell
- Supports 300+ of your favorite CLI tools
- Autocomplete for cd, git, docker, npm, yarn, kubectl, brew, aws and hundreds more!
> All cloud features are opt-in. Data is encrypted at rest.
The most generous interpretation is this is misleading (I would consider it outright lie) given they currently also have telemetry which isn't even opt-out, it's always on. See last time this was discussed:
It uses some sort of substring/edit distance matching and LRU heuristics, and it works almost every time. Sometimes it feels like it's reading my mind.
How do they open those dialogs for completing commands on them? Those menus don't seem to be text based, but part of the terminal window GUI.
I imagine it's quite a bit of engineering effort to make that not janky. I wonder how hard it is to make cross platform. Maybe other OSes have easier hooks? Easier permissions for sure.
As soon as I thought "hey a cool-ish terminal app" I was already sure that it would be macOS only. So I searched on the landing page, documentation and FAQ but couldn't find any info about platform support, which gave me some hope.
Only once I subscribed to the mailing list and got to the download page, I was told that Linux and Windows users are on a "waiting list".
I'm going to refrain from using any app that settles on macOS only as a first target, even if they plan to support other platforms in "the future".
Like the ecosystem wasn't already a walled garden without deliberately excluding viable platforms for no reason other than being "cool".
If you use Python, Rich and Textualize is leading the forth. I am currently using it to build a frontend to a DRF backend and I am really enjoying the experience.
The terminal auto-completion is nice looking. I want to like this, however for me I'm not seeing enough value to justify having a third party potentially snooping my terminal.
Maybe that’s too critical, but can’t see myself migrating away from iterm2 for somerhing proprietary.
And I think it's an interesting idea to move the autocomplete out of band where it doesn't write to the same window where my work is (if I were on a mac), especially if it's still keyboard driven as they say.
But I still wouldn't touch this for other reasons.
I don't see any justification for this being proprietary.
There are already decades old autocoplete in most shells and the enhacement concept that this brings is just not that special that it deserves putting up with having to install proprietary binaries, being limited to their mac-only platform support, tolerating their data collection, and of course whatever their licensing arrangement is or will be.
It's not a bad enhancement idea, it just seems like it's barely a little spiff-up over ordinary bash completions.
Ordinary completions are already a flexible system that you can customize to do quite fancy things if you wanted. They already execute commands to generate output. It's a trivial addition, conceptually, to feed that output into any of the many xdialog-alikes instead of directly to stderr. This is a little slicker that literally just that, but not by much.
And the existing shells are already on all platforms not just mac, and fully open source, and don't have any telemetry or client/server architecture, and so are the xdialog-alikes, and so are the autocomplete definitions.
Is not typing `-u` really worth installing a whole app that spies on you? Especially since shells like `fish` (FOSS!) will have you just write `-u` + up arrow?
This seems objectively worse to me but I guess we'll see whether they find a real market.
Questions about 'sign up' for a cli type app, I mean, I get it, but also slightly uncomfortable with that.
Love to try it out, if others have seen this issue and can hint at a resolution I'm all ears.
There's also WSL2 which gives you a full-blown Ubuntu 20.04 running inside Windows.
Thanks but thanks, I have ctrl + R
I dug the code a bit, didn't find it out.
Though…if they make it simple to get a refund on the ‘Microsoft tax’ you have to pay whenever you buy a new computer but have no interest in running windows past making sure the thing actually works I might have a tiny, tiny bit of sympathy for them.
For a terminal that wants to send telemetry.
I suppose the waitlist is only up for Europe?
Does this also require to log in, like the other one, warp (not the HTTP library)?
Don't think I'll do that.
It's a terminal!
https://news.ycombinator.com/item?id=27277819 - Launch HN: Fig (YC S20)
oh they haven't removed the tracking thing in the url: https://fig.io/?ref=hn
i would love to see what's the algorithm of hackernews, do they give a little boost when it is about their products?
It's so refreshing to see a product try to reinvent something so core to developers everywhere. Really excited to see where this goes!