Live data from Hacker News

Microsoft eyes $10B bet on ChatGPT

semafor.com

31–40 of 541 posts

Re: Microsoft eyes $10B bet on ChatGPT

#31

It’s really impressive watching OpenAI go from “no one can use this model because it’s too dangerous” to “everyone can use the model online but you can’t have the model because we’re selling it to Microsoft for billions of dollars.”

What is their sincere interpretation of "Open" in "OpenAI"? Their models are not distributed publicly, correct? Do they describe the network characteristics and training data openly? Where's the "open" part? EDIT: I just read their charter and it describes "benefit of humanity" (laudable if feasible for them to deliver on it) but nothing about openness.

Pure historical legacy. Today they are Open to all monetization avenues.

Re: Microsoft eyes $10B bet on ChatGPT

#32
post #2

I guess monetizing chatGPT should be simple, I would pay a monthly subscription for it in a heartbeat. Are there any competitors at this point that can match its performance? How hard would it be to replicate?

That’s my question - is it that it’s a huge advance that’s revolutionary or is it an incremental innovation that can be replicated from current state of the art with some effort.

It can be replicated, it's basically clever fine-tuning of GPT-3 with human assistance. Google has model that can do what ChatGPT can for some time already.

Re: Microsoft eyes $10B bet on ChatGPT

#33

Correct me if I’m wrong, but this is a big bet on a company that just had their lunch eaten on another product by open source, no? Didn’t see much of Dall-E after Stable Diffusion release

There are open source / OS model-based projects competing with OpenAI’s text generation. At least one, NovelAI, also has a Stable Diffusion fork incorporated into their paid product.

At least so far none of them are close to what’s been achieved with ChatGPT, and that could in itself be driving some of the hype. ChatGPT was probably more impressive to me on the first impression being familiar with the limitations of existing AI text gen projects.

Re: Microsoft eyes $10B bet on ChatGPT

#34

It’s really impressive watching OpenAI go from “no one can use this model because it’s too dangerous” to “everyone can use the model online but you can’t have the model because we’re selling it to Microsoft for billions of dollars.”

What is their sincere interpretation of "Open" in "OpenAI"? Their models are not distributed publicly, correct? Do they describe the network characteristics and training data openly? Where's the "open" part? EDIT: I just read their charter and it describes "benefit of humanity" (laudable if feasible for them to deliver on it) but nothing about openness.

It was always just marketing.

Re: Microsoft eyes $10B bet on ChatGPT

#37
post #2

I guess monetizing chatGPT should be simple, I would pay a monthly subscription for it in a heartbeat. Are there any competitors at this point that can match its performance? How hard would it be to replicate?

They give access to text-davinci-003 and code-davinci-002 models through their API which are strongly related.

But they have done massive amounts of refinement and continue to do so I believe. But I think the amount the refinement has helped improve may not be as much as you would think. For example I think a big part of the refinement is the guardrails.

Re: Microsoft eyes $10B bet on ChatGPT

#38

Earlier quoted context omitted.

ChatGPT Caused 'Code Red' at Google: https://www.cnet.com/tech/services-and-software/chatgpt-caus...

ChatGPT is one of those few technologies that: 1. Competes against and destroys Google's primary profit center (search). 2. Google cannot copy, improve, and give away for free to use as a moat without severe collateral damage to itself. There's very little incentive to use a search engine to find an ad ridden website to answer a question when one can ask ChatGPT directly for a much better response. Google is rightful…

Surely Bing’s ChatGPT interface will have ads. There’s no money in giving it away for free. And Google can embed a clone of it in their search interface as well. So I don’t see this being a competitive advantage for anyone against Google. If anything, it’s one more moat for new search providers to cross.

Re: Microsoft eyes $10B bet on ChatGPT

#39

Earlier quoted context omitted.

ChatGPT Caused 'Code Red' at Google: https://www.cnet.com/tech/services-and-software/chatgpt-caus...

ChatGPT is one of those few technologies that: 1. Competes against and destroys Google's primary profit center (search). 2. Google cannot copy, improve, and give away for free to use as a moat without severe collateral damage to itself. There's very little incentive to use a search engine to find an ad ridden website to answer a question when one can ask ChatGPT directly for a much better response. Google is rightful…

At this point, I think the problem is the validity of the response from GPT though. I’ve seen some terribly unpredictable and incorrect results. At least with Google I have options to pick through.

Once that is solved, it’s gonna be hard to go back to picking through the increasingly ad and spam ridden always changing UI that is Google.

Re: Microsoft eyes $10B bet on ChatGPT

#40

Earlier quoted context omitted.

ChatGPT is one of those few technologies that: 1. Competes against and destroys Google's primary profit center (search). 2. Google cannot copy, improve, and give away for free to use as a moat without severe collateral damage to itself. There's very little incentive to use a search engine to find an ad ridden website to answer a question when one can ask ChatGPT directly for a much better response. Google is rightful…

Surely Bing’s ChatGPT interface will have ads. There’s no money in giving it away for free. And Google can embed a clone of it in their search interface as well. So I don’t see this being a competitive advantage for anyone against Google. If anything, it’s one more moat for new search providers to cross.

Eroding your competitors business model is a business tactic. Many business run as loss leaders.
Post reply on HN