Researchers tested the technology and discovered that it was broken

Unblocking Visible Watermarks Using Artificial Intelligence: a Case Study of a Novel by Fizi and Feizi

This August, researchers at the University of California, Santa Barbara and Carnegie Mellon coauthored another paper outlining similar findings, after conducting their own experimental attacks. It indicates that all invisible watermarks are vulnerable. This newest study goes even further. Some researchers had hope that visible watermarks may be developed to survive attacks, but Feizi and his fellow researchers say that even this more promising type can be manipulated.

So, it was really amazing the extent to which having an AI creative partner helped unblock me. Even though it was still… The plot of this book was what I was trying to figure out. It would be pointless for a reader to consume a novel that was 100% generated by an artificial intelligence, without a single human touch. I have no idea what that is doing.

The AI Moment: Why are We Getting More GPUs? How Do We Go about Building a Portfolio of AI Models? A Post-Theoretical Analysis

We’ve arrived now at the nature of art, so I’m going to make a hard shift to GPUs. This is what I mean about Kevin — we can go everywhere with Kevin. I just want to make sure we hit it all.

People make art. The AI moment has provided us the opportunity to ask that question in a serious way. Because the internet has basically been like, “To make money.” And I think there’s a divergence there, as our distribution channels get flooded. I am not sure whether we will hit the answer in the next 10 minutes.

So, the last time you and I spoke, you said something to me that I have been thinking about ever since. This man controls the entire GPU budget at Microsoft — every dollar that flows into GPUs, right here.

It’s easier now than when we talked last time. So we were in a moment where I think the demand… Because a bunch of AI technology had ripped onto the scene in a surprising way, and demand was far exceeding the supply of GPU capacity that the whole ecosystem could produce. That is resolving. We’re getting better each week and it’s great that we’ve got more good news ahead of us than bad. It makes the job of judging these gnarly conflicts easier.

This week there was some reporting. In The information, you pointed out that Microsoft is invested in smaller models that don’t require as much compute. Are you able to bring the cost of computation down over time?

I think we are. When you bill an application for something, you end up using a full portfolio model, that’s what I’m going to say here. So, you definitely want to have access to the big models, but for a whole bunch of reasons. If you can offload some of the work that the AI application needs to do to smaller models, you probably are going to want to do it.

And some of the motivations could be cost. There could be something going on. Some of them could be that you want to run part of the application locally because you don’t want to transit sensitive information to the cloud. There’s just a whole bunch of reasons why you want the flexibility to architect things where you’ve got a portfolio of these models.

Source: Microsoft CTO Kevin Scott on Bing’s quest to beat Google and the future of AI art

OpenAI API Embedding: OpenAI is the Key to Building an AI Application on a Cloud Platform for Collaboration, Collaboration, and Software Development

Well, let me deploy my finest press training and say that if you are an API customer right now — like you’re using the Azure OpenAI API or using OpenAI’s instance of the API — you don’t have to think about what the underlying hardware looks like. It’s an API. It is presented to you to be the simplest possible way to go build an AI application on top of that API.

I’m looking at Copilot in Office 365. $30 is the price for a seat. That price is insane. I think some people are going to think it’s very valuable, but that’s not a massive market for an AI pricing scheme. Can you bring that down?

I think we can bring the cost of the artificial intelligence down. OpenAI reduced the cost of access to the GPT-3.5 API for developers by a factor of 10 this spring. That was almost entirely covering a lot of performance improvements. So, the chips are getting better price performance-wise, generation over generation. And the software techniques that we’re using to optimize the models are bringing tons of performance without compromise to quality down. And then there are some other techniques of how do you compose the application of smaller and larger models that help. The cost goes down. And the price is just what value you’re creating for people. So the market sort of sets the price. And if the market tells us that the price for these things is too high, then the price goes down.

We are getting good signals about the price. The thing you just said is important to me. It’s very early for commercialization of generative artificial intelligence. So you have a bunch of things you need to figure out in parallel. What is the market for those things and how do you price them? There is no reason to overprice things. The thing that you want is everybody getting value from them, as many as humanly possible. Over time, we will figure that out.

When I think about compute — these big models, running tools for customers — obviously, the story there is Nvidia chips, right? H 100s is access to. Building capacity is what it is. They’ve got 80 percent of the overall market share. How much do they make you feel?

Source: Microsoft CTO Kevin Scott on Bing’s quest to beat Google and the future of AI art

AMD, AWS and the CUDA platform: I’m excited to see what you can do in next generation graphics processing chip development (Extended Abstract)

They are one of our most important partners. The relationship between us and them is good and we work with them on a daily basis.

They make their own chips, I look at Amazon. I talked to the CEO of AWS a few weeks ago on Decoder. He didn’t sound thrilled that he had a dependency on a chipmaker. They want to build their own systems. Are you looking into the possibility of custom chips? Are you thinking about diversifying that supply chain for yourself?

Competition is a very good thing when you want to make sure that the prices you pay for your products are not too different from the price you paid for something else. I know Lisa Su, from AMD, is here at the conference. We do a lot of work with Lisa, we think they’re going to become more and more important in the marketplace in the future, I think they’re making increasingly compelling graphics card offerings that are more and more important. I think there’s been a bunch of leaks about first-party silicon that Microsoft is building. We had been building it for a long time. So.

I’m not confirming anything. I will say that we have had a pretty substantial investment in Silicon for a long period of time. We want to make sure that we use the options we have available to us in order to build these systems. Over the past several years Nvidia has been the best choice. They have been.

Is that because of the processing power in the chip, or is it because of the CUDA platform? Because what I’ve heard from folks, what I heard from Lisa yesterday, is that actually, what we need to do is optimize one level higher. We need to optimize at the level of PyTorch or training or inference. And CUDA is not the thing, and that’s what Nvidia’s perceived mode is. Are you in agreement with that? Is that the case that you are dependent on the chip? Do you depend on their software infrastructure? Are you working above that level?

Well, I think the industry at large benefits a lot from CUDA, which they’ve been investing in for a while. So if your business is like, “I got a whole bunch of different models, and I need to performance tune all of them,” the PyTorch-CUDA combo is pretty essential. There are not a lot of models that we’re working on.

There are other tools that let you do exactly what you said and help with the development of high-resolution kernels for your system, such as the open-sourced tool Triton, which was developed by OpenAI. Even if it is just a graphics card manufacturer, you still want to make it easy to deploy multiple different hardware types in production, so you can easily adjust across all of them.

So I asked Lisa yesterday, “How easy would it be for Microsoft to just switch from the Nvidia to AMD?” And she told me, “You should ask Kevin that question.” You are here. How easy right now would it be if you needed to switch to AMD? Are you working with them on anything? And how easy would it be in the future?

So yeah, it isn’t trivial to mess with this hardware. It is all big investments. If you build your application that way, you don’t need to care. And there are a bunch of people who are not building on top of these APIs where they do have to care. And then, that’s a choice for all of them individually about how difficult they think it might be. The only thing that the customer sees is the application programming interface, and that is part of the big complicated software stack.

I think the open- source stuff is going to help a lot of people. We opensourced the model called Phi that is currently in demand on Hugging Face. A bunch of open-source innovations we’re excited about. But I think the big models will continue to make really amazing progress for years to come.

I don’t know if it is worth thinking about the models as moats. So there are some things that we’ve done, and a path forward for the power of these models as platforms, that are just super capital intensive. And I don’t think even if you’ve got a whole bunch of breakthroughs on the software, they don’t become less capital intensive. So, whether it’s Microsoft or someone else, the thing that will have to happen with all of that capital intensity… because it’s largely about hardware and not just software, and it’s not just about what you can put on your desktop — is you have to have very large clusters of hardware to train these models. It’s hard to get scale by just sort of fragmenting a bunch of independent software efforts.

Source: Microsoft CTO Kevin Scott on Bing’s quest to beat Google and the future of AI art

A question for Kevin Scott on Bing’s quest to beat Google and the future of AI art: (II): Cryptographic Watermarking Technologies

I’ve got a few more questions. Please begin lining up if you have questions for Kevin. I’d love to hear from all of you. I want to make sure we are careful about what we say about authenticity and what kind of thing it is. In the past I talked about a lot. There’s a lot of ideas about how you might mark content as real or mark it as generated by AI. We’re going to see some from Adobe later today, for sure. Have you made any progress here?

Yeah, I think we have. One of the things I think we talked about before is for the past handful of years, we’ve been building a set of cryptographic watermarking technologies and trying to work with both content producers and tool makers to see how it is we can get those cryptographic watermarks — they’re manifests that say, “This piece of content was created in this way by this entity” — and have that watermark cryptographically preserved with the content as it gets moved through transcoders and CDNs and as you’re mashing it up a bunch of different ways.

Text is definitely harder. There are some things that you can do in the generation of the text to subtly add a statistical indicator to how you are generating the text. It is harder than visual content in that it can be very easy to hide a watermark in the picture and not change the experience you have when viewing it. It is a tougher problem for sure.

There’s nothing I want more than someone sending me an email that says it was generated from AI at the bottom. My inbox would be repaired by thinking about it.

I know what my preference is for those emails. I will tell Cortana to remove them immediately. It’s a good warning to all of you. It is gone if you write me a message.

Source: Microsoft CTO Kevin Scott on Bing’s quest to beat Google and the future of AI art

Kevin Scott on Bing’s quest to beat Google and the future of AI art [Microsoft CTO Kevin Scott]: Does it need a sensory consumer in medicine or health?

Pam Dillon: Good morning, Kevin. Pam Dillon of Preferabli. This question is not being generated by ChatGPT. We have been talking a lot about assimilating the world’s knowledge. It is likely that there will be specialized bodies of knowledge where there is real domain expertise. Do you mean that in medicine or health there needs a sensory consumer?

Kevin Scott: Yeah, we are thinking a lot about that. And I think there’s some interesting stuff here on the research front that shows that those expert contributions that you can make to the model’s training data, particularly in this step called reinforcement learning from human feedback, can really substantially improve the quality of the model in that domain of expertise. A lot of thought has been put into the medical applications.

So one of my direct reports, Peter Lee, who runs Microsoft Research and who’s also a fellow at the American Medical Association, wrote a great book about medicine and GPT-4, and there’s a whole bunch of good work. All of that is what you said. It is how — through reinforcement learning, through very careful prompt engineering, through selection of training data — you can get a model to be very high performing in a particular domain. Over time, I think we’re going to see more and more of that with a lot of different domains. It is really exciting.

Source: Microsoft CTO Kevin Scott on Bing’s quest to beat Google and the future of AI art

What’s going on with the black keys? Part I: On the case of the Black Keys as a source of income for artists

Alex: Hi Kevin, my name is Alex. I have a question about my past. Yesterday, the CEO of Warner Music Group, Robert Kyncl, was talking about his expectation that artists are going to get paid for work that is generated off of their original IP. Provenance isn’t given by LLMs today. My question to you is from a technical standpoint: Let’s say that somebody asks to write a song that’s sort of in the style of Led Zeppelin and Bruno Mars. The Black Keys are music that kind of sound like Led Zeppelin and the LLM is currently using them. Is it possible that the Black Keys music was used in generating the output so that the artist could get compensated in the future?

Ks: Yeah, maybe. I don’t think it’s a good thing for human writers to be asking that particular thing. I know there was this big lawsuit with Ed Sheeran about exactly this, where it’s pretty easy for a human songwriter to be influenced in very subtle ways. And a lot of pop songs, for instance, have a lot of harmonic similarity with one another.

You have to consider both sides of the issue. What is actual, AI aside, how do you measure the contribution of one thing to another? Which is hard. And then technically, if we were able to do that part of the analysis, you probably could figure out some technical solutions. If you make sure that you are not having generations that are parroting, then it is very easy to do. It can either be in whole or snippets. It is more difficult to figure out how a given generation has been influenced by the huge amount of data that any piece of data has.

Gretchen Tibbits: I am Hi, and I have a DC Advisory with me. Rewind slightly from the question the gentleman just asked. There have been some questions from publishers about the information that has been used to train these models. Forget about generating music and the next, but that’s been trained and asking for percentages or rights or recognition of that. I’m wondering — and not asking you to comment on any active case — but philosophically, thoughts on that?

Source: Microsoft CTO Kevin Scott on Bing’s quest to beat Google and the future of AI art

KS: Is Moby Dick about a whale? How will you remember him? Comment on Kevin Scott on Bing to beat Google and the future of AI art

KS: So, here’s a thought exercise. By raise of hands, how many of you have read Moby Dick? So, I’m guessing that all of you who raised your hand probably read Moby Dick many, many years ago — high school, college maybe. And if I ask you, you could tell me Moby Dick is about a whale. There’s a captain. Maybe you remember his name is Ahab. Maybe he has a fixation issue with this object. You could tell me more about the man. Some of you might be able to recite a couple of passages from the book exactly as they appear in it.

The person says thatKS The thing will get sorted out. It depends on judges and lawmakers to figure this out, so we won’t know the answer until we figure it out as a society. The thing the models are trying to do isn’t… They aren’t a huge repository of all of this content. You’re attempting to build something that, like your brain, can remember conceptually some of these things about a thing that was present in the training. We will have to see.

Source: Microsoft CTO Kevin Scott on Bing’s quest to beat Google and the future of AI art

Jay Peters: What is going on in the epoch of science and what is going to happen next? The balance of trade and what was going on

I don’t want to see any of my readers lose their minds… I am an author and I do not want to see anyone lose their mind as well. There are incentives for people to make content and earn money writing books. Especially, God forbid, folks who sit down and do the work of writing a really thoughtful, super well-researched piece of nonfiction. Someone who is passionate about writing a piece of fiction. They should be paid for it. This is a new way of doing things with content. And I think we still have some big questions to ask and answer about exactly what’s going on and what is the fair way to compensate people for what’s going on here.

What is the balance of trade like? Because hopefully, what we’re doing is building things that will create all sorts of amazing new ways for creative people to do what they’re best at, which is creating wonderful things that other people will consume that creates connection and enhances this thing that makes us human.

Jay Peters. Hi, Jay Peters for The Verge. I thought of this thing when you said you did not want to read spammy garbage, and Microsoft published an article that recommended a food bank as a travel destination. And that was made apparently in combination with algorithmic techniques, techniques with human review. If something is bad with human intervention, how can we trust the summaries?

KS: Yeah. The things that were involved were more about how the human pieces of them were working. Honestly, that would’ve been a little bit better if there’d been more AI.

KS: No, I’m not blaming anyone. I think the diagnosis of that problem is some of these things on MSN — and I know this is true for other places — gets generated in really complicated ways. There was a Columbia trained journalist who was sitting down writing this, and all of a sudden there was a faulty tool that was doing the things that they used to do. That’s not what was going on here.

KS: I think you will judge the content by its quality. You are the ultimate arbiter of, “Is this good or bad?” if it is directed at you. Is it true or false? One of the things that these tools may be useful for is helping navigate a world where there will be a large number of tools that can produce low-quality content. And having your own personal editor-in-chief that’s helping you assemble what you think are high-quality, truthful, reliable sources of information and helping you sort of walk through this ocean of information and identify those things will be, I think, super useful. I think what you all are doing, by the way — and many of you in the room, I’m sure, are in media businesses — I think having all of this content out there makes your job more important.

KS: Way more important. Because somebody has to have someone that they trust, that has high editorial standards, and who are helping figure out signal and noise. It’s absolutely true.

Source: Microsoft CTO Kevin Scott on Bing’s quest to beat Google and the future of AI art

How Do I Get My Aims Right? AI-Generated Content, and Why I’m Trying to Write a Science Fiction Book

Correct. And I actually agree with that. But the point that I was making is the useful thing about the tool is it helped keep me in flow state. So I’ve written a nonfiction book. I have never written a novel before. The useful thing for it wasn’t actually producing the content, but helping me get unstuck, like if I had an ever-present writing partner or an editor with infinite amounts of time to spend with me. It’s like, “Okay, I don’t know how to name this character. Let me tell you what they are about. Give me some fun names.”

That is the model we have today. We’re in the context of the writers strike resolving. Even in that conversation, they were not worried about the model’s capabilities today. There will be a GPT-5 and GPT-6), right?

I’m pretty sure that you are going to want to use some of the tools to help produce content. When we were playing with this stuff for the first time, I thought about writing a science fiction book since I was a teenager. And I started to attempt doing that with GPT-4, and it was terrible at using it in the way that you would expect. You cannot just say that you have an outline for a science fiction book you want to write. Chapter one should be written.

So, there’s nothing about an AI being 100 percent of that interaction that seems interesting to me. I don’t know why I would want to be consuming a bunch of AI-generated content versus things that you are producing.

Sometimes, AI-generated content is good, and sometimes, it’s not. I don’t think it’s very interesting. It’s kind of a technical problem, whether or not you’re ingesting things into your training process that are causing the performance of a trained model to become worse over time. That is a technical thing. I think it’s an entirely solvable problem.

We’ve got an increasingly good set of ways, at least on the model training side, to make sure that you’re not ingesting low-quality content, and you’re sort of recursively getting—

The flip side of that is you also make a lot of tools that can create AI content. And you see these distribution platforms immediately being flooded with AI content. A search engine that floods it’s own artificial intelligence with too much can lead to things like model collapse and a reduction in quality. How do you find out what is in there?

So, I think that’s one of the opportunities that we can have right now in the conversation about how these AI agents are going to show up in the world. It isn’t necessarily preserving exactly what the funnel looks like, but it is transparent about the mechanics of it in order to make sure you are aware of what is going on. and you no longer know how to viably run your business.

Now, I think the compensation structure and how things work just evolves really rapidly. And it feels to me like, even independent of AI, things are changing very rapidly right now — like how people find an audience for the things that they’re creating, how people turn audience engagement into a real business model. It is difficult because some of the funnels are hard to work out. You don’t really know what’s going on in the algorithm that’s directing traffic to your site

I don’t think anyone wants that. It’s certainly not the thing that I want, individually. There needs to be a healthy economic engine where people are all participating. They’re creating stuff, and they’re getting compensated for what they create.

Source: Microsoft CTO Kevin Scott on Bing’s quest to beat Google and the future of AI art

What Should I Write About a Phone? [Review Note: AI-Powered Search Platforms] Are Going to be Fair, Isn’t It?

If an AI search product can just summarize for you what I wrote in a review of the new phone, why would I ever be incentivized to create another review of a phone if no one’s ever going to visit me directly?

You are researching on how to wring out the cables in a house that you are remodeling, like you are planning a vacation. That may involve purchasing some things or spending some time reading a pretty long thing because you can’t get the information that you need in just some small transaction that you’re having with an agent. I think it’s unclear the extent to which the dynamic will actually change. I think the particular thing is everybody is worried about referrals, and how is this going to… If the bot is giving you all the answers, what happens to referral traffic?

Yeah. So I think what you want from a search engine and what you’re going to want from an agent is a little more complicated than just asking a question and getting an answer. A whole bunch of the time, what you want is you’re trying to accomplish a task, and asking questions are part of the task, but sometimes, it’s just the beginning. Sometimes it is in the middle.

I think the conventional wisdom is that [in] an AI-powered search experience, you ask the computer a question, it just tells you a smart answer, or it goes out and talks to other AI systems that sort of collect an answer for you. That is the future. I think if you just broadly ask people, “What should search do?” You get an answer when you ask a question. That changes the idea of how the web works. The fundamental incentive structure on the web is appearing in search results. Have you thought about that with Bing?

Yeah, broadly. We need to be asking ourselves how everyone can participate, and what’s fair. At the end of the day, that is the goal. We’re all creating these big platforms, whether it’s search as a platform, these cloud platforms… we’re building AI platforms right now. I think everybody is very reasonable in wanting to make sure that they can use these platforms in a fair way to do awesome work.

And I think the only thing that anybody can ask for is that you do high-quality product work, and you want marketplaces to be fair so you can compete. It is true for large companies and small companies that are trying to break through. Just whatever it is that is that notion of fairness is what everybody’s asking for. And it’s a complicated thing to go sort out. I will not comment on what’s going on on the East Coast right now.

Source: Microsoft CTO Kevin Scott on Bing’s quest to beat Google and the future of AI art

Microsoft and the AI renaissance: Where do we stand? What are we doing now, and where do we go? What do we need to do to improve?

I think Bing is a good search engine. It’s the search engine that I use. I have been at Microsoft for over six and a half years. When I got there, I was still an avid user of the search engine. It got to a point where the Edge browser and Bing search combo was more than enough to be my driver browser-plus-Edge combo. And we’ve seen a growth in market share.

I could talk about anything with Kevin. He’s a maker. You are a renaissance. We were talking about crimping ethernet cables before we walked out onstage — literally anything. We have to discuss the topic of artificial intelligence. So I want to just ask from the beginning: Microsoft kicked off a huge moment in the AI rush with the announcement of Bing, the integration of OpenAI into the products. There’s obviously Copilots. What is happening with that? Is the use of Bing increasing as a result of the integration of AI into it?

I also asked Kevin some pretty philosophical questions about AI: why would you write a song or a book when AI is out there making custom content for other people? Well, it’s because Kevin thinks the AI is still “terrible” at it for now, as Kevin found out firsthand. But he also thinks that creating is just what people do, and AI will help more people become more creative. This conversation got deeper, I really enjoy talking to Kevin.

I co-hosted the Code Conference last week, and today’s episode is one of my favorite conversations from the show: Microsoft CTO and EVP of AI Kevin Scott. Kevin and I love talking about technology together, so if you caughtKevin on Decoder, you know that we enjoy talking about technology together. I really appreciate that he thinks about the relationship between technology and culture as much as we do at The Verge, and it was great to add the energy from the live Code audience to that dynamic.

Kevin and I talked about how things are going with Bing and Microsoft’s AI efforts in general now that the initial hype has subsided — I really wanted to know if Bing was actually stealing users from Google.

Kevin also controls the entire GPU budget at Microsoft, and access to GPUs is a hot topic across the AI world right now — especially access to Nvidia’s H100 GPU, which is what so many of the best AI models run on. Microsoft itself runs on H100s, but Kevin is keenly aware of that dependency, and while he wouldn’t confirm any rumors about Microsoft developing its own AI chips right now, he did say a switch from Nvidia to AMD or other chip vendors should be seamless for Microsoft’s customers if the company ever does make that leap.