Nice work! Been playing with Langchain and was not aware of patterns.app.
This whole space is moving so fast its hard to keep up for someone whos immediate day job doesn't revolve around this space. Congrats.
Show HN: AskHN
21–30 of 139 posts
Re: Show HN: AskHN
#22Am I correct in understanding that this doesn't actually generate answers based on HN, but instead finds semantically-near comments and sends them verbatim to GPT to summarize? Seems like a good enough hack, though I'd love a detailed writeup of how to actually specialize an existing LLM with additional training data (like HN).
Agreed, I think the better approach is to do some custom tuning but that becomes cost prohibitive very quickly. Not really much different than Algolia with a minor GPT-3 integration but neat project regardless.
Re: Show HN: AskHN
#23It seems to write in the generic "style" of GPT, instead of in the style I would recognise as a HN poster. Is that because of something baked into how the training process works? It lacks a sort of casualness or air of superiority ;)
There was no training process, this is just running GPT with relevant HN comments as part of the prompt.
If he wanted it to replicate that classic HN feel he would either have to extend the prompt with additional examples or, better yet, use finetuning.
I guess he could also just randomly sprinkle in some terms like 'stochastic parrot' and find a way to shoehorn Tesla FSD into every conversation about AI.
Re: Show HN: AskHN
#24Is there a way to opt out of one's comments being used for this?
Why would you want to? Genuinely wondering.
I for one am oh so proud that my valuable ramblings contributed to this majestic machinery.
Re: Show HN: AskHN
#25Earlier quoted context omitted.
I am guessing they already were? But this is 100% pure, concentrated HN not contaminated with nonsense from the rest of the web :)
If it's really trained exclusively off of HN comments, I expect most of the bot's responses will evade the actual question but spend several paragraphs debating the factual specifics of every possible related tangential point, followed by an thinly-veiled insult questioning the user's true motivations.
If you're not arguing over the semantics, rather than OP's clear-enough intent, are you really on HN?
Re: Show HN: AskHN
#26Earlier quoted context omitted.
I am guessing they already were? But this is 100% pure, concentrated HN not contaminated with nonsense from the rest of the web :)
Is it exclusively HN comments and nothing else? How does a model like that know how to speak English (noun/verb and all that) if you are starting from scratch and feeding it nothing but HN comments?
In a nut shell, this is using openai’s api to generate embeddings for top comments on hn, then also generating an embedding for the search term. It then can find the closest related comments for the given question by comparing the embeddings and then send the actual text to GPT3 to summarize. It’s a pretty clever way to do it.
Re: Show HN: AskHN
#27Is there a way to opt out of one's comments being used for this?
Nah, it's no big a deal, its not like cambridge analytica will happen again. They're just using your data to train AI. Who knows may be based on the way you comment, you may get suggestions on which medication you need, or if it's time for the Redbull/starbucks coffee. Nah, all is good. Nothing bad will happen in allowing companies to scrape comments and build models. They're very ethical.
In fact, people here are suddenly not so concerned that the model is not open. There is no oversight on how data is being used
They are just proud to get answers from a text generator.
Re: Show HN: AskHN
#28Is there a way to opt out of one's comments being used for this?
I agree: when I signed in, I never agreed to let anybody use what I write to do anything they want ! I only agreed to let everybody read, understand, interact with what I wrote
Actually, it makes me feel as bad as knowing that CAPTCHA were used to train image recognition models...
I think it could be a good time to reconsider the question of the consent. I may agree that my words are used to train some IA... but 1) I must be asked (kindly) first and 2) it won't be free!!! (it may be paid to me or the service provider like HN... but it's NOT unpaid work ;-) )
Re: Show HN: AskHN
#29Is there a way to opt out of one's comments being used for this?
Nah, it's no big a deal, its not like cambridge analytica will happen again. They're just using your data to train AI. Who knows may be based on the way you comment, you may get suggestions on which medication you need, or if it's time for the Redbull/starbucks coffee. Nah, all is good. Nothing bad will happen in allowing companies to scrape comments and build models. They're very ethical. In fact, people here are su…
The BIG DEAL is not THAT specific instance but the fact that the ML crowd think it's OK to take everything without even asking permission
Re: Show HN: AskHN
#30[dead]