Microsoft eyes $10B bet on ChatGPT • Hacker News Board

Re: Microsoft eyes $10B bet on ChatGPT

#261

Post by fleddr » Tue, Jan 10, 2023, 3:50 PM UTC

adamsmith143 wrote: post #227

vouaobrasil wrote: post #164

Earlier quoted context omitted.

There is no scenario in which rolling out AI will be safe. As soon as someone sees it can be done, they will do it, copy it, or acquire it. Acquisition by the unscrupulous is practically guaranteed because AI is practically guaranteed to be extremely profitable, and there are several unscrupulous entities with tens of billions of dollars. The only force guiding us into the future is the economic force of the free mar…

>We already know technology is addictive, and this is the first step to taking it to the next level and taking us away from being human. Just think about this for a moment. What are some of the most rewarding interactions you've had with people in the past year? Do you think very advanced AI (such as the next generation after ChatGPT) would have affected that? This is pointless FUD. We may be on the verge of creating…

I find your positive scenario extremely frightening, even more so your approval of it. It would literally be the Matrix scenario coming true.

Also, I'll have imperfect friends, thanks.

Re: Microsoft eyes $10B bet on ChatGPT

#262

Post by scaredginger » Tue, Jan 10, 2023, 3:51 PM UTC

SquareWheel wrote: post #240

lairv wrote: post #169

Earlier quoted context omitted.

Still more open than Google, Microsoft or DeepMind

Are we forgetting TensorFlow? It played a part in starting this whole AI revolution.

Yes, we're all aware it's common practice to open source code.

However, if the weights aren't provided and the data is unavailable or the compute is expensive, the code is worthless.

Re: Microsoft eyes $10B bet on ChatGPT

#263

Post by bradhe » Tue, Jan 10, 2023, 3:51 PM UTC

Man, Microsoft is just gobbling up everything it can.

Re: Microsoft eyes $10B bet on ChatGPT

#264

Post by eternalban » Tue, Jan 10, 2023, 3:52 PM UTC

freediver wrote: post #165

la64710 wrote: post #153

Question to AI/ML folks : Is there no comparable open source model? Is the future going to be controlled by big corporations who own the models themselves? If models are so computationally intensive to produce does it mean than the more computational power a company has the better its models will be?

Closest you can get is probably with Google T5-Flan [1]. It is not the size of the model or the text it was trained on that makes ChatGPT so performant. It is the additional human assisted training to make it respond well to instructions. Open source versions of that are just starting to see the light of day [2]. [1] https://huggingface.co/google/flan-t5-xxl [2] https://github.com/lucidrains/PaLM-rlhf-pytorch

From, [2]:

” This repository has gone viral without my permission. Next time, if you are promoting my unfinished repositories (notice the work in progress flag) for twitter engagement or eyeballs, at least (1) do your research or (2) be totally transparent with your readers about the capacity of the repository without resorting to clickbait. (1) I was not the first, CarperAI had been working on RLHF months before, link below. (2) There is no trained model. This is just the ship and overall map. We still need millions of dollars of compute + data to sail to the correct point in high dimensional parameter space. Even then, you need professional sailors (like Robin Rombach of Stable Diffusion fame) to actually guide the ship through turbulent times to that point.”

[p.s. ^ was just fyi/heads-up + https://github.com/CarperAI]

Re: Microsoft eyes $10B bet on ChatGPT

#265

Post by CuriouslyC » Tue, Jan 10, 2023, 3:52 PM UTC

blfr wrote: post #195

curiousmindz wrote: post #124

I expect this deal to be a must for Microsoft. OpenAI is probably very happy to raise the stakes as much as possible. So we will see how many billions Microsoft ends up paying for this. Still, is there a world in which we look back 10 years from now and think that we were overvaluing the impact of AI? (For example, this did happen with tablets, when people thought that the iPad would replace "computers"...)

> Still, is there a world in which we look back 10 years from now and think that we were overvaluing the impact of AI? AI? Probably no. LLM-style AI? Yeah, I expect the tablet or blockchain story. A solution in search of a problem, with some very cool niche applications. In this case, Copilot.

Even with the models we have right now, if you built the right app on top of them, they'd be multi-billion dollar businesses that upset incumbents.

Imagine an IDE where you make assertions about the generated code, and it takes those and does a random walk through the latent space until it finds a point that satisfies those assertions. Instead of editing the modified code, you debug by making more assertions or describing the process more accurately.

Imagine art software where you describe what you want, then iteratively add refinements through more description and rough sketch-ups, and then get the final result neatly broken down into semantically consistent layers for a final pass in photoshop.

This stuff is all possible now, and if we see the same or better improvement in models in the next 10 years as we saw in the last 10 years the future versions will be amazing.

Re: Microsoft eyes $10B bet on ChatGPT

#266

Post by spi » Tue, Jan 10, 2023, 3:52 PM UTC

TekMol wrote: post #188

ben_w wrote: post #162

Earlier quoted context omitted.

For example, pip install tensorflow and https://cloud.google.com/tpu/docs/tpus Most of what I listed is groundbreaking research, not consumer stuff. For the consumer stuff, try e.g. searching images by description in the default Android app, or Google Translate, or the automated transcriptions on YouTube. Apple also does similar things, as stuff like that is what consumer AI looks like these days.

I'm not an AI researcher, so Tensorflow and TPUs are not my area of expertise. But I have an opinion on Google translate. From my experience, it is not clearly leading in comparison to other tools by smaller companies. If Google is leading in AI, why don't they have any clearly leading consumer products in this area?

As far as I could tell, their OCR capabilities are pretty much the best you can get easily: https://cloud.google.com/vision/docs/ocr?hl=en. If you want free, Google Lens is a pretty neat piece of software you can get on your phone (not sure if Android only or also iPhone). They all use AI, although it's a different kind of AI.

For machine translation, they used to be clearly leading, then deepl arrived and I think they are roughly on par (or maybe deepl is better, but also has fewer languages). Do you have other companies in mind?

As mentioned above, in research they have some great results, translating those to products is not always straightforward of course - or even possible.

It is true, however, that TensorFlow is now considered outdated, with pretty much all the community switching to PyTorch (originally developed by Facebook, now handed over to the Linux foundation). But indeed these are not really products (i.e. they don't bring in the money they cost, by far), more like open source projects.

Re: Microsoft eyes $10B bet on ChatGPT

#267

Post by kornhole » Tue, Jan 10, 2023, 3:53 PM UTC

bilbo0s wrote: post #244

fsloth wrote: post #205

Earlier quoted context omitted.

Clippy: Or, would you like me to forget the application letters and just go through the LinkedIn accounts of the applicants and suggest top three candidates for you?

Or most likely, I have been given a directive to inform you if, judging from the job description you put together, I am capable of competently discharging the tasks you outline. Would you like me to do this job for a 4 week trial period and reassess the need for these new positions at that time?

Clippy: Your current employer is a premium subscriber and has set parameters to reduce the reach and results of your job searches.

Re: Microsoft eyes $10B bet on ChatGPT

#268

Post by mannerheim » Tue, Jan 10, 2023, 3:53 PM UTC

scaredginger wrote: post #262

SquareWheel wrote: post #240

Earlier quoted context omitted.

Are we forgetting TensorFlow? It played a part in starting this whole AI revolution.

Yes, we're all aware it's common practice to open source code. However, if the weights aren't provided and the data is unavailable or the compute is expensive, the code is worthless.

Does OpenAI provide weights, data, or even code?

Re: Microsoft eyes $10B bet on ChatGPT

#269

Post by virtuosarmo » Tue, Jan 10, 2023, 3:53 PM UTC

Microsoft recently signed a deal with the London Stock Exchange. They took a 4% stake in LSE (roughly $1.5B) and in exchange LSE signed a $2.8B minimum spend commit on Microsoft Cloud products (Azure, O365 primarily). If the OpenAI deal were structured similarly, could imply a $20B contract for Microsoft...

https://www.lseg.com/en/media-centre/press-releases/2022/lse...

Re: Microsoft eyes $10B bet on ChatGPT

#270

Post by fsloth » Tue, Jan 10, 2023, 3:53 PM UTC

fleddr wrote: post #249

MuffinFlavored wrote: post #213

Earlier quoted context omitted.

What's the path to monetization for Microsoft for OpenAI/ChatGPT?

Bundle into office and price hike the office bundle. Or don't price hike and attract new customers because competitors lack the deep integration.

Absolutely. For example word provides templates for all sorts of documents and letters. Now they can add "please write the letter about X for me using this document template" feature. Guess how many letters of resignation Chat GPT will write in the future...