Back to our roots | Hacker News Reader

Back to our roots

honnibal.dev

138 pointssaeedesmaili10 comments

The bit that sticks out in the story is this:

Justin and Sebastián left the company with the transition to investment

Obviously a ton of context is missing but if you are planning to stake the future of your company on a product that had until now been developed by these two gentlemen why weren't they going to be part of that future?

One thing I have grown an appreciation for over the years is the power of very small teams, i.e groups of less than ~4 or so. When you have a very small number of very capable people you can paper over a lot of deficiencies.

Overall though a very realistic view into what it's like trying to scale up a startup. If it's any consolation most of them blow up just like this, don't feel bad if this is a pattern you recognise from your past - it's just how it is. The game is hard and failure is expected, scale ups are by far the most vulnerable time in a companies history and yet you need multiple of them to "make it" and each one is completely different from the last.

> When you have a very small number of very capable people you can paper over a lot of deficiencies.

That surprises me. I would assume that the fewer people you have, the less talent you have to cover all needed skills. But maybe you mean something else with "deficiencies"?

I am a fan of small, committed, and accountable teams. Unfortunately, sometimes the management layer is stifling them. I recently worked on a team of four devs where we had... 9 (nine!) managers in every meeting. None of them were able to help us get specs, architecture reviews/approvals, or new environments. For nine months.

I'd call this a team of 13

wow what a candid, and humble take! Really impressive. Lots to be learned here.

However, one thing I am curious about is how their VC investor, SignalFire, is ok with this "back to it's roots" terminus. Do they still own the same amount of the company as before?

There is some info on Honnibal's Twitter (https://x.com/honnibal/status/1813650728222880157):

Delip Rao: Curious how do you go back from VC funded to bootstrapped? Did you return money to investors? And were they okay with just that? Regardless, I am happy you are doing more lib dev. You have very good taste.

Matthew Honnibal: There's no change in ownership, it's just that we don't have any more VC money to spend --- so the operating reality is like it was before. We think it's helpful to explain this.

Delip Rao: So if you consult and do other stuff as Explosion to bring money in, you have to pay distribution of the profits to the shareholders on your cap table?

Matthew Honnibal: If we ever pay a dividend, all shareholders will get some, (and the VCs have liquidation preferences that mean they'd take a larger share than their ownership until the money is paid back). But we're some distance away from dividends -- we'd pay ourselves a salary from consulting

Delip Rao: That's right. I am surprised your VCs were okay with this arrangement. Usually, they nudge the company towards an acquihire in such situations.

The title is a little misleading. They raised $6 million from SignalFire (who also invested in Grammarly) for 5% of the company [1]. As SignalFire retains that shareholding, they are not less independent (nor more) than that have been in the past couple of years. They are just not looking for more investment and so are more "independent-minded".

I used SpaCy quite expensively pre-2020. It still has a lot of great uses, and more predictable than LLM models. But now many NLP tasks that seemed near impossible, or required a lot of expensive annotated training data (which their product Prodigy is used for) can now be coded in very little time with LLMs.

[1] https://explosion.ai/blog/weve-sold-5-percent-of-explosion

Yeah I was hoping there will be some bit about how LLMs make many of the tasks that people used SpaCy for a lot easier and cheaper. That is a bigger threat to the project going forward I don’t see how it can exist in the same form as today.

> For most difficult decisions you’ll never know what would have happened if you had done things differently.

This sentence really resonates with me and should live rent free in a lot of people's mind.

Reply on news.ycombinator.com