Hacker News Reader

Comment by tomlue | Hacker News Reader

I'm a fan of OpenAI, but this is nonsense. All of human existence is mostly other people's ideas. Among a ridiculously huge list of other things, OpenAI benefits from the mountains of labor that made scalable neural networks possible.

Have they had their own good ideas? Definitely. Are they benefitting of ideas mostly invented elsewhere? Also Definitely, just like everybody else.

View on HN

I'm not really a fan of openAI, but I think we're seeing the classic mistake of confusing product with technology. Steve Jobs / Apple didn't anything you'd call new ideas either (obvious cliche but so is the criticism). It's execution and design once the tech reaches a certain level

mostly agree. Though I wouldn't underestimate the tech and engineering work behind OpenAI. That microsoft partnership is no joke.

I've heard Altman (on the Lex Friedman podcast) and Sundar Pichai (on the Hard Fork podcast) say things to this degree. The thing that OpenAI really managed to crack was building a great product in ChatGPT and finding a good product market fit for LLMs.

Well sure, but there still aren’t any other LLMs at the level of GPT3/3.5 let alone GPT-4. GPT3.5 just using the API returns fantastic results even without the ChatGPT interface (which isn’t terribly hard to replicate, and others have using the API).

There are dozens if not hundreds of companies that could’ve done something profound like ChatGPT if they had full access to GPT3/3.5. And honestly, OpenAI stumbled a lot with ChatGPT losing history access, showing other users’ history… but that doesn’t matter much as the underlying technology is so profound. I think this really is a case of the under-the-hood capability (GPT3/3.5/4) mattering more than productization and execution.

(Now I think there are not a ton of companies that could do what Microsoft is trying to do by expanding GPT4 to power office productivity… that is a separate thing and probably only about 3 companies could do that, at best: Microsoft, Apple, and Google… and theoretically Meta but their lack of follow through with making Metaverse useful makes me doubt it.)

Hm, I wonder how much of the API’s performance is related to training/finetuning done by OpenAI planned towards the ChatGPT product. I think the RLHF is partly product design and partly engineering.

What's the product market fit for LLMs, and how does OpenAI fill it?

How much money is openai making from chatgpt premium now? How much revenue are they they making from the api?

It’s a good comparison. And once again tech enthusiasts are confused and outraged that the product people are getting credit for tech they didn’t invent. Once again missing the forest that people buy products, not tech.

There’s an awful lot of judgement, engineering and technique that goes into a really well thought out product. It’s often deeply underestimated, and culture makes a huge difference in execution. Bing/Sydney came out after ChatGPT, based on exactly the same tech, but was hot garbage.

Bing/Sydney is better than ChatGPT. It had serious bugs in beta testing

It’s a dramatically worse chat bot, but being able to search the internet does give it an additional useful capability, while limiting it to five interactions papers over its psychotic tendencies.

Try it again - seems very good now (and allows longer conversations too).

It’s interesting to see the ad implementation, I recall some predictions that Microsoft would be particularly apt at finding a way to integrate advertising organically. Instead it just seems to have made the bot more stupid because sponsored products are forced into its recommendations.

I’ve gone back to just using the GPT API, unless I absolutely need to search the internet or information after 2021 for some reason.

I don’t know, my family talks almost daily about how amazing bing chat is. The Sydney eta was kind of crazy, but the core product seems to be doing well.

It is definitely solving a different problem than chatgpt though, and maybe a less inspiring problem. Chatgpt is like an open world game where you can do anything; Bing chat is just a point solution for the vicious spiral of SEO and Google’s profit motive that rendered search results and web pages so useless.

There are three modes to Bing Chat, and at least two models involved. AFAIK only the "creative" mode uses GPT-4 (the other two some flavors of GPT-3.5). As far as I can see Bing/Creative is no different to ChatGPT+ (the non-free version) which is also GPT-4 based.

Rather than execution or design perhaps this time it was mainly about being unethical enough to sell out to investors and unflinchingly gather enough data in one place by carelessly ignoring copyright, authors' desires, privacy regulations, etc.

I don't think you understood me. I'm with you, but given that often OpenAI is accused of using public ideas for profit, this, in turn, means they are obviously not in a dominant position. So far they are just better than the others.

> OpenAI benefits from the mountains of labor ...

Can't you say the same about Google? Google lives from the labor of others.

Not entering into the debate if this is ethically or not just saying that OpenAI is not much different. When photo services such as Google Photos recognize the Eiffer Tower in Paris they are using images from others.

Reply on news.ycombinator.com