Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Launch HN: Bloop (YC S21) – Code Search with GPT-4

news.ycombinator.com

Post reply on HN
Print view View original

61–70 of 131 posts

Re: Launch HN: Bloop (YC S21) – Code Search with GPT-4

#61

Post by s1k3s »

tkiolp4 wrote: post #46

I don’t understand these new saas that rely on GPT. It surely is my lack of understanding, but since the interface to GPT is mainly text: why any company would pay for N different GPT saas (e.g., one for searching for code, another for searching for documentation on confluence, another for allowing your business analysts with their queries, etc.)? Wouldn’t it make more sense for any company to have only ONE interface…

Investors want AI. Do you see how crappy their landing pages look? This one is pretty decent compared to what we've seen showcased here in the past months. It's as if someone at YC told them to ship it asap, with no consideration for anything other than "slap AI on it".

Re: Launch HN: Bloop (YC S21) – Code Search with GPT-4

#62

Post by cjonas »

intalentive wrote: post #56

Where to go next? A Toolformer that calls an AST parser. In VS Code I am running Copilot and IntelliSense side by side. Two competing tab completions -- ridiculous. They should be integrated. I want a Copilot that listens to the interpreter / compiler, so that it only makes correct suggestions.

I can't stand how co-pilot hijacks the standard code completion, often overriding the typescript completion I actually want or generating something distracting.

There should be a way to only run completions when prompted.

Please upvote this issue if you run into the same problems: https://github.com/community/community/discussions/9817

Re: Launch HN: Bloop (YC S21) – Code Search with GPT-4

#63

Post by renewiltord »

evrydayhustling wrote: post #57
renewiltord wrote: post #39

Very interesting. I want you to know that > ‘Private’ here means that no code is shared with us or OpenAI at index time, and when a search is made only relevant code snippets are shared to generate the response. (This is more or less the same data usage as Copilot). is the reason I went from "ha, cool tool" to "okay, let me go download it". Quite surprised that you don't actually have a login wall to download. Missin…

I'm interested why this made a difference to you. If you are concerned about sharing code with Bloop/OpenAI, isn't sharing code during use as bad as sharing it at index time? The more you use it, the more of your code will be exposed...

Only code in the search space is. If it were an "index in the cloud" product, one mistake and I lose the codebase. If it is a "query exposes a portion", a mistake in where I use it exposes a piece.

If I can contain the blast radius of my errors my error budget is more available.

Re: Launch HN: Bloop (YC S21) – Code Search with GPT-4

#64

Post by nico »

eddsh1994 wrote: post #53
hn_throwaway_99 wrote: post #49

Earlier quoted context omitted.

The point is that the interface to Bloop (or whatever 3rd party SaaS product) is just plain text. In programming parlance I would call this a facade [1] on top of GPT-4. The benefit is that you, the end user, can just continue typing things like "Where do we create our database pool?" in a search box, but under the covers Bloop can use GPT-4, but then eventually upgrade to GPT-5, GPT-6, or some other much better curr…

Agreed that may be the case but all OpenAI needs to do is release a nice UI for GPT and they've suddenly pushed these SaaS out of the market. Why fund miners over the companies making pickaxes?

That will probably happen. But it will take some time. Twitter did that. At first they were super open let everyone build apps, then they started blocking and copying the most popular ones, or acquiring them.

However, the companies will probably soon be able to switch out the LLM they are using, so potentially they could use Google’s PaLM, or Facebook’s Llama, or some open source LLM, who knows.

They can also fine-tune the models for just-your-data, so that the responses for you are better than just using a general LLM.

Re: Launch HN: Bloop (YC S21) – Code Search with GPT-4

#65

Post by tedsanders »

tkiolp4 wrote: post #46

I don’t understand these new saas that rely on GPT. It surely is my lack of understanding, but since the interface to GPT is mainly text: why any company would pay for N different GPT saas (e.g., one for searching for code, another for searching for documentation on confluence, another for allowing your business analysts with their queries, etc.)? Wouldn’t it make more sense for any company to have only ONE interface…

To me, your question seems to boil down to the economics of vertical integration.

Two thoughts:

First, GPT is all marginal cost, no fixed cost. So the marginal economics of 1 super app vs 10 specialized mini apps is roughly the same.

Second, imagine the same argument applied to a utility like electricity. "Why would any company pay for lightbulbs from one company, HVAC from a second company, and appliances from a third company?"

Early on, electricity was complicated and you might actually purchase all of these from the Edison Illuminating Company. But there are returns to specialization, and today the winning formula is different companies that specialize in each. The company best at manufacturing lightbulbs is not necessarily the company that is best at manufacturing washing machines.

Similarly, you could ask: "Why would any company pay for a laptop from one company, an operating system from a second company, and office apps from a third company?"

In some cases it may make sense to vertically integrate (e.g., Apple is happy to sell you a combined laptop + OS + Pages/Numbers), but in many cases the specialized players still do fine (e.g., you might buy a Dell laptop, a Microsoft OS, and Notion/Sheets).

So I think it's very much an open question as to whether the winning approach will be a Swiss army knife (single product) or a toolkit (multiple products).

Re: Launch HN: Bloop (YC S21) – Code Search with GPT-4

#66

Post by nico »

intalentive wrote: post #56

Where to go next? A Toolformer that calls an AST parser. In VS Code I am running Copilot and IntelliSense side by side. Two competing tab completions -- ridiculous. They should be integrated. I want a Copilot that listens to the interpreter / compiler, so that it only makes correct suggestions.

> I want a Copilot that listens to the interpreter / compiler, so that it only makes correct suggestions.

And that builds a test first. Then iterates on the new code against the test until it passes.

ChatTDD?

Re: Launch HN: Bloop (YC S21) – Code Search with GPT-4

#68

Post by louiskw »

swyx wrote: post #2

congrats on launching! i honestly feel like a bad user for this but i have yet to adopt a semantic search engine for code for some reason, respite codeium and sourcegraph also offering more advanced code search thingies. any ideas on how to break force of habit?

That's the million $ question. Autocomplete (Copilot) is install and forget. We don't have that luxury, search has to give you that 'aha' moment the first time you use it.

For one, having the engine respond in natural language makes a big difference. The last generation of semantic code search (we built one) used transformers to retrieve code, but as a user you'd still have to read through an entire code chunk to find the answer.

Also, LLMs (probably starting with GPT-4) can now reason properly. The capability to make a search, read the results and execute an entirely new search based on its reasoning skills, and do this iteratively, until it finds the answer, is a huge jump from just using semantic search on its own.

Re: Launch HN: Bloop (YC S21) – Code Search with GPT-4

#69

Post by vanjajaja1 »

eddsh1994 wrote: post #53
hn_throwaway_99 wrote: post #49

Earlier quoted context omitted.

The point is that the interface to Bloop (or whatever 3rd party SaaS product) is just plain text. In programming parlance I would call this a facade [1] on top of GPT-4. The benefit is that you, the end user, can just continue typing things like "Where do we create our database pool?" in a search box, but under the covers Bloop can use GPT-4, but then eventually upgrade to GPT-5, GPT-6, or some other much better curr…

Agreed that may be the case but all OpenAI needs to do is release a nice UI for GPT and they've suddenly pushed these SaaS out of the market. Why fund miners over the companies making pickaxes?

I think the key is that OpenAI wants to focus on building AGI and not waste time appeasing users requests for UX shortcuts for GPT powered code search. Sure, they can hire an extra team to manage the user's requests for UX, but at that point they're moving away from their target goal.

Re: Launch HN: Bloop (YC S21) – Code Search with GPT-4

#70

Post by ggordonhall »

elanzini wrote: post #67

Congrats on the launch! If multiple pieces of code from different files are being referenced in the response, it would be nice to have clickable refs that take you to that piece of code in the repo.

Yes that's a great point. At the moment we in the 'Results' section we display the code snippets returned by the semantic search engine. These don't always line up with the code that GPT is referring to because we include as much of the file that the snippet is from in the prompt.

We're working on a new interface to make this clearer, where we'll display the lines of code that GPT has referenced in its answer.

Post reply on HN