Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Fig is sunsetting (fig.io)
148 points by patrikcsak on March 12, 2024 | hide | past | favorite | 66 comments


The related acquisition thread a few months ago: https://news.ycombinator.com/item?id=37296401

Edit: And the Launch HN from March 2021:

https://news.ycombinator.com/item?id=27277819


It’s so interesting to see the full circle of acquisition PR.


Nice. These guys started up, cashed in, sold out, and bro'd down all in the span of three years.


You’d be a lot less snarky if it was you getting rich.


"Fig team will be joining Amazon Web Services (AWS) and Amazon has acquired Fig's technology!" does not have the ring of "getting rich". Seems like acquihire + asset sale, quite possibly for a negligible price, and with liquidation preferences common stock is likely worthless. Not enough information to say for certain but deals along these lines are far more common than meaningful exits.


Why would they sell if it's for a negligible price? Seems like a pretty terrible deal if all they get out of it is to work at Amazon.


Lots of reasons...

* some return, even if negligible, for investors and the feeling of having done right by them

* decent hiring bonuses - so, some financial compensation even if the founders' equity was worthless

* reputation: they get to claim a successful exit, whatever the reality (so many founders dispense advice based on "multiple exits" which were all fire sales)

* continued life for the product... although in most cases the product is shut down and the team ends up splitting up and working on the acquirer's existing products

There's also just the psychology of it. After failing to raise the next round, you start a process to find some home for yourselves / your team / your product. Having gone through that process, if you come up with any offers, you're pretty likely to take the best one even if a purely rational analysis would suggest that you'd be better off doing something completely different.


* stress relieve: moving from running a company, worrying whether you can pay the bills next month either for the company or for yourself to a job with a good income can remove a lot of stress from one’s personal life.


Perhaps because it was a better deal than any of the alternatives. I left a startup a few months before it was (as far as I know) acquihired, partly because I could see the end of the runway approaching. All of the extant team were offered roles at the acquiring company, which I'm sure they and their families appreciated. Is "all they get out of it is to work at Amazon" preferable to a Big $$$ Exit? Maybe not. But the latter may not have been an option IRL. As a side note, I've been working at Amazon for nearly ten years now, and I have not been displeased with the improvement in my financial situation thereby.


Job at Amazon + investors get x% back is better than no job and 0% for investors …


sorry, but this feels like projection.


See https://ourincrediblejourney.tumblr.com for tons of examples

Lots of fun "there will be no changes or interruptions to the service" followed next month with "we have decided to shut down the service" stories https://ourincrediblejourney.tumblr.com/post/658027964213215...


Having contributed to the Fig autocomplete specs, I find this sad. The Amazon product Fig was built into basically works as replacement, which is good. Still, the core value of this product are the open-source autocomplete specs: https://github.com/withfig/autocomplete. What's going to happen to that? It looks like they are still using it in the Amazon product. It should definitely be possible for an open-source re-implementation of the Fig UI to use those specs. There is a lot of knowledge encoded in there!


The open-source project will continue to be maintained! The open-source autocomplete specs are used by the CodeWhisperer CLI and will be for the foreseeable future.


Where can I read more about the space of FOSS, not-mining-your-data-for-profit, don't-call-home, code completion solutions?


Thanks Matt, best of luck to the team!


withfig/autocomplete is also used by other projects, who might be in a position to fork if something happens. The ones I know of are:

- https://github.com/microsoft/inshellisense

- https://github.com/warpdotdev/Warp


oh thank you for sharing this! I knew about warp but didn't realize it was using the autocomplete specs


That's sad but expected. I actually had switched and disabled fig a few days ago because the app was using a lot of resources and probably had some memory leaks and crashed.

I find the autocomplete feature useful. I don't trust the GenAI commands even with gh copilot. I really hope that one day we can get a tool that will simulate the consequences of a command on your particular machine/environment but I admit that would be very hard.


Natural language to command line is the kind of functionality people dream about but it seems incredibly dangerous. How can I trust it to do what I intend every time? ChatGPT certainly isn't good enough as I've learned trying to use it on files I thankfully had a backup of.


You can't trust it.

You need the same (and probably more) robust controls in place as you would use for a more junior developer.

A human will be influenced by the potential outcome of any proposed solution. For example they might worry about the impact to their employment status if they are wrong, while an LLM will just spit out an answer that is "most likely correct".


Fig was one of the first dev tools that I have been so happy to be able to use every day and would gladly pay for. I’m very sad about this.


Funny how "sunsetting" became a politically correct newspeak for "failed"


Being acquired by Amazon has nothing of "failed" in my book.


In the YC book it failed for sure. Startups are not "lifestyle business" in the YC definition.


It is not political nor failure.

It is corporate speak for we cashed out and the new owners bought us to strip us to parts so they can get what they want out of us (i.e: talent, patent/intellectual propety) while dominating the market and reducing (potential) competition.

It is a feature, not a bug for the shareholders.

Capitalism working as intended.


Did it really fail though?


Is Hyper next? Makes me a bit worried about Pydantic too, and a several other open source dev tools (like ruff, and all HTTP GUI clients) that have taken VC money.


Are there alternatives to fig for linux? Seems like amazon code whisperer is macos only.


Combination of zsh completions with fzf picker and gh copilot for CLIs.

I have a good setup here that you can learn from - https://github.com/fluxninja/dotfiles


Not a direct answer to your question but

> We have several exciting new CodeWhisperer for command line features coming in 2024 including support for Linux


Not explicitly what you're asking for but fish shell gets you a long way into what fig was offering for usability improvements, but without being a closed source SAAS cloud app laden with telemetry.


Many people suggesting zsh, but if you’re willing to move to fish you’ll get many features out of the box.


Zsh + powerlevel10k + zsh autocomplete


Sunsetting before I even knew about it.


An interesting pivot but I guess helping devs sync on multiple computers was never going to be a big market.


True. There's https://chezmoi.io/ for this, but no commercial market for it.


This is at least the second thing call Fig that is shutting down. There was also a video game crowdfunding website called Fig - https://www.wikiwand.com/en/Fig_(company)


Also docker compose before it became docker compose

https://orchardup.github.io/fig.sh/


> CodeWhisperer for command line does not include features similar to Fig Scripts, Dotfiles, Plugins, or Servers and we do not plan to support these features any time soon.

It's rare to see this level of honesty and straight-forwardness. Thanks for that.


Who the hell comes up with these words like "sunsetting"? I was working with a dental company and their software we integrated with our system stopped working. I called the company and asked why and they said "We sunset this product". It took me a while to understand what they were saying then they asked for money because they released a new one. Sunsetting was just a fancy word for a money grab.

Wouldn't be easier to say we abandon the project or kill it? Does sunset sound fancier?


I don’t believe it’s a new term. It means they are intentionally phasing out or terminating something.


In software world it is killing it so they can sell a new one in general.


really enjoyed using fig last year but it was linked to my old company account and now i cant get access as im no longer there. tried to add this amazon implementation but it just seems like an elephant compared to fig. litterally just want the autocomplete power and fuzzy search lookup for branches was so nice. anyone know any anythink lightweight like fig used to be let me know


when will you, personally, learn to not to bother getting invested in products made by people trying to sell their company to a FAANGM?


Read about codewisperer just now (linked from this article). Anyone know how it compares to the warp terminal’s AI features?


I am not surprised. I tried to sign up for Fig after the news it is acquired by Amazon, the sign up page in the app is closed.

I check the Github issue page and it is stale for several months. Definitely this deal is acquire-hiring and all the Fig workflow will be absorbed to Amazon products.


Anyone have experience with any options for something that's not run by Amazon?


I am curious from someone who uses fig every day - what’s the rundown on its value?


I am/was a daily user. To me it was just very good autocomplete on terminal/CLI commands.


Autocomplete for CLI apps, gui for directory changes, just quality of life things.


Hmm I'm curious. What does "gui" mean? It yanked you out of the command line in a browser or finder?

And was this autocomplete server based or happened locally and privately?


It added a correctly positioned autocomplete list in front of the cursor in almost any terminal app.

The core server was local, its two most used features were directory, and CLI arguments autocomplete.


Autocomplete of S3 buckets.

E.g

aws s3 ls s3://<magic…>

Also brief documentation of any arguments or flags for the command you are typing.


Hahaha, I knew it. We all knew it.

Seeing the aftermath of corporate acquisitions, especially in the tech industry, is always interesting. Over my ten years in the field, I’ve seen three companies go through big changes during and after being acquired. It’s funny how companies try to convince everyone that nothing will change, even though it almost always does. People’s reactions range from hopeful to in denial about the changes.

Is it because people are delusional or psychopaths trying to make other people believe their lies.



I may be dense, but I don't see how Fig fits into the analogy of the five monkeys getting sprayed by icy water.


> Is it because people are delusional

Dunno how any dev could rely on some SaaS pay2win autocomplete or the SaaS batch job processor stuff they sold.

Or rely on these kind of companies in the first place.

I mean, sure, pad your resume by making your employer use Best Practice SaaS products, but using Fig seem to be a "don't shit where you eat", messing with your workflow.

edit: Heh reading some launch thread from 3 years ago is hillarous.

https://news.ycombinator.com/item?id=31222277

I wonder how many of the yay sayers were shills?


what a shocking news


I don’t care


They raised at least 2.2M in funding. Then they got acquired. Now it's over. I do think it's time we stop pretending that you can build a fully-fledged company on top of a developer tool or plugin.

Lots of companies like this are going to be struggling in the next 2 years.


At $2.2M in funding, they are (likely) on a SAFE. There's no board, meaning no real VC pressure. They used that money to build... it's not like they were a beloved, widely used developer tool ruined by VC money.

It sucks, but to build certain things, you sometimes need money. $2.2M isn't a ton – they raised in 2020, so that's roughly only (market) salaries for 3 people for those 3 years.

Not everything will work out. But good on Fig for trying, and good on Fig for finding a solution that (hopefully?) did right by themselves and their team, and good on them for leaving Fig open-sourced.

(And to the Fig-ians reading this, congrats :) )


Yeah you're probably right about there not being any pressure. the 2.2M was a seed round. 2.2M is a lot for 3 people salaries for 3 years in my opinion


With taxes, office space, laptops, offsites, insurance, etc a rough rule of thumb is $250k for an engineer. $250k * 3 = $750k, and over 3 years that's $2.25M (doesn't include software, hosting, lawyers, contractors, marketing, etc).

Yeah, they probably spent less than that (esp the founders salaries), but my point is that $2.2M isn't as much money as it seems.


Apparently Fig was a terminal enhancement that provided a fancy autocomplete system.


Well, who would thought, acquired a few months ago and poof




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: