You can't compare the accuracy of speech recognition to LLM task completion rates. A nearly-there yet incomplete solution to a Github issue is still valuable to an engineer who knows how to debug it.
Sure, and no doubt people paying for speech recognition 25 years ago were finding uses for it too. It depends on your use case.
A 13% success rate is both wildly impressive and also WAY below the level where I would personally find something like this useful. I can't even see reaching for a tool that I knew would fail 90% of the time, unless I was desperate and out of ideas.
I disagree. I think about this a bit as having a developer intern, on whom I can't rely to take much of a workload, and definitely nothing on the critical path, but I could say to them "Take a look at these particular well-defined tasks on the backlog and see which ones you could make some progress on" - I feel there's good value in that.
And the nice thing about an AI here is that I think it will actually find a different subset of these tasks to be easy than a human would.
Yeah, but a developer intern already has human-level AGI to support the on-the-job developer training you are going to help give them. Any LLM available today, or probably in next 5-10 years for that matter, has neither AGI nor the ability to learn on the job.
My experience of working with interns, or low-skill developers, is that the benefit normally flows one way. You are taking time out from completing the project to help them learn. Someone/something of low capability isn't going to be relieving you of the large or complex tasks that would actually be useful, and be a time saver - they are going to try to do the small/simple tasks you could have breezed through, and suck up a lot of your time having to find out and explain to them how they messed up. Of course Devin doesn't even have online learning, so he'd be making the same mistakes over and over.
> A nearly-there yet incomplete solution to a Github issue is still valuable to an engineer who knows how to debug it.
Not sure if I can agree. There would definitely be a value in looking at what libraries the solution uses, but otherwise it may be easier to write it oneself, especially when the mistakes are not humanlike.
One factor that no one seems to have noticed is that both cofounders, Ravi and Joshua, have already built fairly big companies. Ravi was the co-founder at Heap (he took over my co-founder spot after I left), and Joshua was the CTO at Benchling.
Both of those companies have raised >$100M. After that success, the prospect of continuing to run a middling company in the shadow of Retool isn't very appealing.
Sure, but wouldn’t users be hesitant to use Ravi and Joshua’s products in the future? Businesses depended on Airplane and even paid for the service so having the rug pulled from under them will leave a bad taste.
Airplane was a startup in a competitive market, it wasn't Google Reader shutting down. Anyone using it was an early adopter and in my experience, most early adopters understand the risk. Maybe it's possible that folks will be hesitant, but after multiple successful exits, maybe that's a hypothetical that won't matter any more, often founders move on to be advisors/board members, and I think any board would be lucky to have Ravi's experience on tap.
Here is another; they fucked over their customers. Didn't even open source it on the way out?! Nobody with a memory for this is going use their products without second thought in the future.
But peoples memories are short and a sucker is born every day.
Would a board be lucky to have Ravi on it? Sure, maybe. So their personal success is secured fuck their users; slow clap.
I'm an airplane user, and have already moved our tooling over to retool. Fuck over their customers? I'm sorry, but you have a choice in what tools you acquire, buy, and invest in, if you don't take into account that the company is small, new, and requires VC funding to exist, I just can't empathize with you if you are angry about it not working out. I use new software because it usually gives me access to the team in a way that lets me build with them and get features that a larger incumbent won't provide, but that doesn't mean it will make the company successful, and if they fail it sucks, but it's not like they are actively trying to hurt me.
There are so many reasons it's not easy or straight forward to open source the software of a startup that failed. I don't know what to tell you, I guess since I've only worked for startups I kind of understand what risks are involved when working with them, but there are also benefits, it's a trade off that you should be making when you're informed.
If you want stability, buy from a profitable and stable company that can pen multi-year contracts that include liability clauses, and if your company has the leverage, get source code/on prem written into your MSA.
All that said: please, don't use new software or tools, avoid startups and steer clear of emerging technology, you'll feel less betrayed.
*I'm not a founder, never have been a founder, and never could be a founder.
> There are so many reasons it's not easy or straight forward to open source the software of a startup that failed
I'm guessing zero effort was made to continue service for existing customers. If there were efforts made that failed, they would have mentioned
it for PR.
> All that said: please, don't use new software or tools, avoid startups and steer clear of emerging technology, you'll feel less betrayed.
Doesn't change what happened.
> *I'm not a founder, never have been a founder, and never could be a founder.
Apparently you don't need to be to have a founder perspective.
Early on, my business relied heavily on Zapier in 2015 and they were just as big as Airplane at that point. If they sold the company and shut it down, I would swear never to do business with the parent company again. Yes I could migrate to something else or just code it if that were to happen, but that is a fuck over.
This is a bit late but yes, I believe you can "fuck over" people with a number of motivations that aren't necessarily "malicious".
For example I may be in a life boat with 5 other people and instead of adhering to rations I eat all the food or drink all the water. Because I can't control myself, I lack empathy, or I'm just incredible self-centered. Whatever the motivations I think many may describe what I did as "fucking over" everybody else.
Indifference can be malicious. Or at least it would qualify for fucking over the customers IMO. At least give them (and the employees) a bit more warning.
> Quite baffling that success is measured in how much you raise.
It's a proxy for size of company, and often one of the two public ones (the other being the number of employees you can count on LinkedIn). It's not perfect, of course; there's always a Zapier which raises 1.3M and then grows sustainably. But for VC funded companies, this lets you compare them, if not perfectly.
I think everyone would agree that revenue, ARR, and profitability are all more important. But those metrics are rarely shared for private companies.
How does one take over a co-founder role? Is it a job title now? If he took over for you he didn’t actually found the company, right?
I’m not asking this sarcastically. Trying to understand how this works because I’ve always taken “co-founder” to mean “I was one of the people that started this company from day 1”
In this context, Ravi who became cofounder of Heap certainly didn't take that route. He was and is fantastic and if anything I'm surprised Airplane wasn't even more successful because of his involvement.
I don't have all the early context, but my understanding (vaguely) is that Heap went through two different incubators, and after one Ravi came on to be cofounder at YC where Heap raised its seed round.
I was employee six and didn't know (though I did know there was a history to Heap before it joined YC and there was another cofounder, I just never knew/didn't dig into who)
There are lots of examples in the use cases section of any low-code vendor, but in short, most internal tools are not in the horizontal categories you listed. They're in vertical, industry-specific categories at SES (Software-Enabled Services) companies.
A quick commerce startup, for instance, is unlikely to find an inventory tracking app off-the-shelf that fits their workflow, since existing ERPs are not going to fit their workflow.
I use them both. I use ChatGPT to do the big parts like the SQL schema and some boilerplate. It works better in Python with fastAPI than Rust with Axum though.
