Great question! So all of the files are generated with either compiles or GPT-3, or if the files already exist on the repo and need to be modified, the bots use GPT-3 code edits to modify the files.

$299/project/month will sound expensive if you compare it to starter kits (usually free), templates (usually < $100), or no-code tools (around $100). But I believe Second will feel very close to having a dedicated developer that can build commodity features for you. I want engineers to use Second for the common stuff, so that they can focus on the stuff that special to the project.

Also, yea pricing is hard! It may change.

> But I believe Second will feel very close to having a dedicated developer that can build commodity features for you.

A “developer” or a “grunt coder”?Or, to use the ~40+ year old terms that even the least trendy places have phased out [0], a programmer/analyst or just a programmer?

[0] Largely because one of the two didn’t really fit into modern development methodologies anymore, since it presumed two narrow of a skillset and responsibility area.

This. Thinking about any of the sizeable apps I develop and maintain, 90% of the job is just thinking of edge cases in the business logic as transmitted to me by the people running the business I'm writing an app for. About 10% is actually coding and testing. While I'm impressed with copilot/GPT's ability to write SQL queries and imitate my coding style to write functions that usually have no obvious bugs, it's the subtle, non-obvious bugs that I'm worried about. Anything it writes demands careful reading and correction - just like anything written by ChatGPT demands a great deal of fact-checking. So much so that it's not really efficient. The idea of reviewing a whole code branch it's written fills me with dread, because who knows where the corner cases are lurking? And when they arise, who knows what else patching them might affect? At least with my own code or code written by humans, I know where things are and can deduce why they were written that way. With copilot, there's no why.

In a sense it's a metaphor for the whole problem of LLMs ... or AI in general. When unlimited compute lets you replace logical deduction with brute force heuristic reduction, it makes it impossible to trace routes and find bugs.

And this boilerplate can't be added fast with Copilot? Which is $100/year.

Maybe. Copilot is more like an extreme programming experience, where you are still piecing together things yourself in the IDE. Second does all of that for you by building complete features and raising a pull request!

I believe that in the future, A.I. bots will be used to build large chunks of the codebase (like Second). And then human engineers will use A.I. bots in the browser to make low level tweaks and additions (Copilot)

Hopefully in your future an AI bot doesn't invent a Second competitor and put you out of business.

Hah. To be honest, 2023 is going to be insane. I'm sure hundreds, if not thousands, of companies are going to emerge which greatly improve the software development process.

I think the software development process is fine, it's all the organizational cruft that slows it down. Your company is a good example of it--you probably didn't use Second to make Second, and being a solo dev had to have been a huge benefit to your velocity.

The things that hurt software development are people/process related, and can't be fixed by software, IMO.

Funny enough, I did use second to build the authentication flows. But I've diverged a bit because I haven't had a chance to fully build out the other modules that Second will need.

I'm planning on rebuilding Second with Second as soon as I have all of the appropriate modules in place!

Please write a blog post about that when you get a chance to do it!

What like scrum and agile? Or can you be more specific? Or point me to blog posts about this stuff.

$299/month sounds pretty reasonable if it does what it says on the tin.

That's $3.6k/year.

You don't need to make up many development hours before you've paid that back.

And it seems like there's good opportunity to stratify billing via modules -- Using only the basic stuff? $. Using more advanced / larger modules? $$.

I agree. I think if it continues to grow it will pay dividends back in development hours saved.

Is it true that the larger the app, the less effective/accurate the bot is? In that case, might it be more sensible to charge a 1-time price?

Good question! I believe the accuracy will still be good, but it will just be more work for the bots to get it right (lots more files and code to reason with). So my OpenAI costs will prob go up a lot for bigger code bases. Hadn't thought about that!

you forgot one alternative - chatgpt, which is $20/month and infinitely flexible

True, but it's still work. Using ChatGPT manually is great, but it's not as fast as configuring a feature and then having the whole thing built and integrated. A lot of users also struggle with knowing what to even ask ChatGPT, unless they are already an experienced developer.

if they aren’t experienced, how will they know the solution is what they need? chicken/egg problem

What would be some examples of features that your service could add to my app?

I could see this used at a big tech companies using golang. I think each company has their startup kit, but it still requires a lot of lifting and lacks many features like open source tools.

yea totally! My plan is to train models on companies design systems as well. I think if done correctly, it could be a huge boost to product development velocity.

Many people implement things like forgot password and billing once and copy/paste in the next projects if they need.

Yea! That kinda works, except that libraries, frameworks, and paradigms evolve so rapidly in the web, often times you need to rebuild it anyways.

I also want to be more supportive of junior developers. Second can help scaffold and get large chunks of standard code in place for them to build on, rather than starting with empty files. This is especially useful when adding a new feature to an existing web application (in which case boilerplates and templates are a no-op)

> libraries, frameworks, and paradigms evolve so rapidly in the web

Speaking of that issue, wasn't GPT-3 trained with data that doesn't go past 2021? Won't this be an issue for any AI assistant using GPT-3 to write code?

I was worried about this as well. However, I think the solution to this is to create fine-tuned models on specific areas that OpenAI models won't be as knowledgable about (like Next.js 13 for example)

I think OpenAI plans to continuously update ChatGPT (aka "gpt-3.5-turbo") but I don't think it has done so yet.

This comment gave me the Dropbox vibes of 2007: https://news.ycombinator.com/item?id=8863 :)

Dropbox also launched in a non-collapsed market. OP has a cool product, but the comment is right - how wide is their moat here?

The biggest lesson to learn from Dropbox is that if you don't build a unique enough product, larger companies will eat your lunch. Microsoft already has Copilot patents, if Second ever threatens them then they have full control over the economics of GPT. Their business is built on the goodwill of a direct competitor.

OpenAI has the GPT models, not Microsoft. Also the open source models like BLOOM are getting better over time and when OpenAI pulls the rug, maybe they'll be competitive.

Fingers crossed. I don't have much hope though. It feels like everyone has said that there's plenty of incentive to knock Nvidia's block off with this AI business, but we've yet to see any real competition.

My guess is that the profitability of ChatGPT is configurable enough to compete with SOTA competitors. Even if a better model crops up, out-pricing OpenAI will be a struggle. It might even require a hosting partnership with AWS or Azure.

> OpenAI has the GPT models, not Microsoft.

Haha no. There is no future where OpenAI operates independently of Microsoft. OpenAI is functionally beholden to MS. 49% is a paper contrivance to present a facade that they're still in control. They're not. The tail is wagging the dog, and in this case the tail is a trillion dollar dragon that OpenAI lowered the drawbridge and invited inside.

Why do you say that?

Because it is true.

At this point, Microsoft has partially acquired OpenAI and is extending their exclusivity deals to use their offerings.

But the only way to compete against them is with open source AI models that are good enough substitutes for OpenAI’s offerings.

With Stable Diffusion being open sourced, it completely ruined monetising options for DALLE-2 and threatened OpenAI to the point where they rushed out ChatGPT.

The same thing happened with GPT-3 and is going to happen again to ChatGPT and we will quickly see an open-source equivalent.

I've been looking around at open source offerings that can run on "affordable" hardware. Right now you can run something like GPT-J on 4 GTX 3090s, which will cost you anywhere between 6k and 8k not including the rest of the PC.

Once functional models get small enough, I think we'll see an absolute explosion in AI as anyone with a laptop will be able to freely tinker with their models.

You can run GPT-J on an M1 or Pi if you want, you'll just have to settle for a pruned model. No self-hosted options (besides maybe Facebook's leaked thing) can stand up to ChatGPT's consistency or size. I've also been playing with this and have had great results for non-interactive use on even the smallest models (125m params, ~2gb RAM). The problem as I see it won't be inference time/acceleration as much as it will be having enough memory to load the model, and settling for lower quality answers. ChatGPT is already pretty delusional, and pruning it doesn't make it any smarter. You can practically feel the missing connections in the quality of response.

So, big takeaway: AI text generation is legible and quick with smaller models, but "intelligence" a-la ChatGPT seems to scale directly with memory.

Could we do two models that are different levels of abstraction? One for "ideas" and one for compressing the idea into words? I've been thinking that smaller specialized networks might boost space efficiency.

I've no clue how these would be wired together, but I have done ideas.

Have you seen the latest Llama and Alpaca models?

The only way any MS decision is not going to go through is if all other shareholders are unanimously against it.

It seems like a stretch to say that would ever be true.

I just released the new version of https//aidev.codes which you can use a prompt kind of like "give me a web page for managing a restaurant" .. it will output it and host it for pennies and if you want a subdomain its $5 per month. It is currently limited to simple static pages. But it can still manage something like recording restaurant schedules in localStorage.

https://news.ycombinator.com/item?id=35086436