How AI Is Fundamentally Reshaping the Web

0:00Today on the AI Daily Brief, the latest legal wrangling between publishers and AI, and before that in the headlines, a lawsuit around AI use in high school. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. To join the conversation, follow the Discord link in our show notes.

0:23Welcome back to the AI Daily Brief Headlines Edition, all the daily AI news you need in around five minutes. One of the breakout use cases of LLMs is, of course, in school. Now, to some, the cynics, the skeptics, or as they would probably describe it, the realists, ChatGPT and its ilk have been nothing more than a cheating tool, a way to get out of having to actually think and write for oneself. For the perhaps more enlightened and future-oriented, ChatGPT represents a new educational aid, something to radically shift how people learn and to provide a new resource for supporting education. Now, what's fairly clear to me is that no matter how long-term optimistic one is around LLMs in school, there's no way that there isn't going to be some bumpiness along the transition.

1:08I personally think that schools are going to have to completely change how they think about education from the ground up, given that these tools now exist, and there's not going to be any good way to actually track who used AI or not. Now, that creates a context for our first story today, where parents of a high school student in Massachusetts have sued their school after the high school accused him of cheating on a paper. The TLDR of what happened is that a student, according to the parents, used AI to assist with the research of a history paper. However, according to those parents, did not use AI to write the paper itself.

1:39The school, however, accused the student of cheating, gave him a bad grade, and even gave him detention. Additionally, he was barred from being inducted into the National Honor Society. Now, the parents are saying this is going to cause irreparable harm, given that the student is currently applying to elite colleges and universities. And ultimately, to me the details of the case, in other words, what the parents are trying to get, and the specifics around the school's lack of clarity around AI use, all of that matters less than the underlying question of this tension in how education is going to happen in the years to come.

2:09Now, this is obviously a fairly dramatic circumstance, given that it's actually getting legal here, but I think it dramatizes the fact that we are in totally new territory where AI is a fact of life that parents, students, and educators are going to have to adapt to in some way, shape, or form. Speaking of adaptation, NVIDIA took a big hit yesterday in the stock market after news broke that the US is considering additional AI chip export restrictions, specifically actually capping exports from NVIDIA and AMD to certain countries. Bloomberg sources tell them that the approach would set a ceiling on export licenses for certain countries, justified by national security concerns.

2:45It sounds like this is not just about China, but also about a number of Arab countries in the Persian Gulf. NVIDIA, which had closed at a record high on Monday, fell as much as 4.2 % yesterday on Tuesday. Part of why this crackdown might have been even more scary to some is that the Gulf states are such an important growth market when it comes to AI. Now, they are also the epicenter of geopolitical tension, given that they sit literally in between the U.S. and China. Bloomberg also notes that in addition to this being about the chips themselves, quote, some U.S. officials have come to view semiconductor export licenses, particularly for NVIDIA chips, as a point of leverage to achieve broader diplomatic goals.

3:20What's more, even if they did decide to move forward with this sort of capping structure, it may not be possible in the time that this current administration has left. All of this, of course, speaks to the geopolitical and national security dimension of AI, which is so fascinating. However, I would be surprised if this moved forward in the next couple of weeks. As the Book of Krusty account on Twitter pointed out, whoever in the White House that mouthed off about looking at chip exports will be looking to walk it back. Last thing Democrats need is teetering markets into election night. Now, speaking of the global nature of AI investment, Blackstone has announced an$8.2 billion investment into data centers into northeastern Spain.

3:55Aragon in Spain is becoming a real destination for these data centers. Blackstone will follow Microsoft and Amazon, who have together so far proposed 19 data center projects to local authorities. Microsoft's planned investment in the region is around 6.7 billion euros, whereas AWS said it would invest 15.7 billion euros. Part of the draw is that Aragon has a large wind power capacity so renewable energy becomes a more viable solution for powering these incredibly power-hungry data centers. Lastly today, another big talent acquisition in favor of AI, the former Chief Information Security Officer of Palantir, Dane Stuckey, has joined OpenAI to lead security.

4:31Stuckey has been at Palantir for the last 10 years, starting in 2014, and given Palantir has extensive experience with government contracts, people are reading this as potentially helping advance OpenAI's ambitions in that domain. Since lifting their ban on selling AI tech to the military last January, OpenAI has worked with the Pentagon on several projects, including ones related to cybersecurity. It also appointed former head of the NSA, General Paul Nakasone, as a board member, which was highly controversial in and of itself. Forbes also added a little texture to OpenAI's ambitions, reporting that they are working with government contractor Kerasoft.

5:03Kerasoft essentially serves as a go-between for government tech procurement working with a wide range of big tech firms. The new partnership will allow OpenAI to offer software and services through the government's central clearinghouse for tech solutions known under the acronym CHESS. CHESS is a contract vehicle that allows the government to purchase services from private industry. OpenAI confirmed they were added to the CHESS contract in May but have not yet contracted with the Department of Defense. Writes TechCrunch, several weeks ago the company posted a job listing for a head of trusted compute and cryptography to lead a new team focused on building secure AI infrastructure.

5:31The speculation is that OpenAI is attempting to build more secure models that can be trained on sensitive material within military databases. So I highly geopolitical headlines this episode, but that is going to do it for the headlines. Next up, the main episode.

6:10you what's okay to say or not. Venice won't patronize you. It simply provides direct access to machine intelligence. No topics are off limits, no ideas are taboo. With Venice, you're in control of the AI as you should be. Pro subscriptions are available for$49 a year or$8 per month. AI Daily Brief listeners receive a 20 % discount on Venice Pro. Visit venice.ai slash nlw and enter the discount code nlwdailybrief. That's nlwdailybrief, all one word. Today's episode is brought to you by Superintelligent. Every single business workflow and function is being remade and reimagined with artificial intelligence.

6:48There is a huge challenge, however, of going from the potential of AI to actually capturing that value. And that gap is what Superintelligent is dedicated to filling. Superintelligent accelerates AI adoption and engagement to help teams actually use AI to increase productivity and drive business value. An interactive AI use case registry gives your company full visibility into how people are using artificial intelligence right now. Pair that with capabilities building content in the form of tutorials, learning paths, and a use case library. And Superintelligent helps people inside your company show how they're getting value out of AI while providing resources for people to put that inspiration into action.

7:25The next three teams that sign up with 100 or more seats are going to get free embedded consulting. That's a process by which our Superintelligent team sits with your organization, figures out the specific use cases that matter most to you and helps actually ensure support for adoption of those use cases to drive real value. Go to bsuper.ai to learn more about this AI enablement network. And now back to the show. Welcome back to the AI Daily Brief. Today, like I said, nominally, the story is about another target coming into the New York Times view, where the Times has told Perplexity basically to stop using their content.

8:02The Times has sent perplexity a cease and desist, demanding that perplexity stop accessing and using its content. The Times, of course, is already embroiled in a lawsuit with OpenAI around using the Times content in their model training. The Times alleged that the way perplexity uses newspaper articles is a violation of copyright. The Wall Street Journal published an excerpt of the cease and desist, which read, perplexity and its business partners have been unjustly enriched by using without authorization the Times expressive, carefully written, and researched and edited journalism without a license.

8:32Now, unsurprisingly, Perplexity, for their part, says that they're taking this seriously, that they're not ignoring the warning, with CEO Aravind Srinivas saying, We are very much interested in working with every single publisher, including the New York Times. We have no interest in being anyone's antagonist here. Notably, Perplexity already has licensing deals set up with a handful of individual publishers, although reports are that their terms are less generous than the eight - and nine-figure licensing arrangements offered by OpenAI. Still, it's a little bit different here than the OpenAI situation, where perplexity is not being accused of using this content to train their models.

9:04In a statement, they said, we aren't scraping data for building foundation models, but rather indexing web pages and surfacing factual content as citations to inform responses when a user asks a question. The law recognizes that no one organization owns the copyright over facts. This is what allows us to have a rich and open information ecosystem, not to mention it gives news organizations the ability to report on topics that were previously covered by another news outlet. So one aspect of this is just the broader media world adapting to the AI landscape. Indeed, it's getting so confusing that the Press Gazette in the UK recently tried to sum up who's suing and who's signing.

9:38On the lawsuit side, Mumsnet, the Center for Investigative Reporting, eight daily newspapers owned by Alden Global Capital, including the New York Daily News and the Chicago Tribune, The Intercept, Raw Story, and Alternet, The New York Times, and Getty Images are all in the midst of lawsuits against OpenAI or other AI companies. Meanwhile, on the signing deal side, there is a big Hearst deal, which was just announced last week, which will include 20 magazine titles and 40 newspapers in total. Microsoft has partnered with FT, Reuters, Axel Springer, Hearst, and USA Today. Condé Nast has signed a multi-year deal with OpenAI, and the list goes on.

10:12Now, interestingly, a lot of these deals aren't about training data, but are about how the content from these publishers shows up in search results. And to me, this gets at the much more interesting part of this story, which is the changing nature of search on the web. For the last 20 years, the default starting point for people interacting with the web has been the Google search bar. You type in what you're looking for or a question you're trying to get answered, and you have a list of little blue links based on Google's indexing. You also have the ability to advertise, to put your link at the head of those search results, and an entire massive industry has grown up around trying to help you get to the top of those search results more organically.

10:49The bargain between publishers and Google has been that Google would be allowed to crawl and index all of these websites in exchange for pointing traffic their way. What's different about the era that we're heading into is that if the AI summary becomes the dominant form of information consumption, there is potentially a lot less of that clicking behavior. In other words, if AI can read websites for you and summarize a bunch of them very quickly, it's immensely time saving for the end user, which is great for them, but it means that they spend less time clicking around. Less time clicking around means fewer sites get visitors and advertisers don't get paid as much.

11:24And this is really what's at stake with perplexity. One of the things that makes perplexity interesting is that it doesn't have the same sort of legacy business model that Google does, which would limit its ability to go whole hog into this new format. And for a lot of people, when it comes to which is the more valuable search experience, it's not even really a question. Dilbert creator Scott Adams tweeted, Try the Perplexity app for five minutes and tell me Google still has a business model. You might never use Google again. And Perplexity is certainly moving quickly. Yesterday, the company announced Perplexity Finance, which includes real-time stock prices, deep dives into a company's financials, the ability to compare multiple companies, studying 13Fs of hedge funds, etc.

12:01Investor Jim O'Shaughnessy writes, Is it just me or does all of the fire Perplexity is bringing should make Google nervous? CEO Arvind Srinivas responded to that saying, Here are the categories perplexity lacks in today relative to Google. Visual images, local maps, navigational latency, needle in the haystack queries, like specific home on real estate or rental markets, stack overflow bugs, link to developer library docs, etc. Shorts, entertainment, shopping, knowledge graphs, and interactive cards. Basically what Srinivas is saying, really more declaring in this post, is that each one of these things is going to be knocked down in turn.

12:34They are, in other words, coming for the king full stop. Now, another interesting thing about Perplexity is that they've been integrating more reasoning models into search, with Srinivas again saying, Perplexity ProSearch is already a research and decision agent, we just don't use the buzzword. Now, how this all resolves in terms of business model remains to be seen. A few weeks ago, the Financial Times reported that Perplexity was starting to have conversations with top brands on a new type of ad model. FT writes, Perplexity is seeking to redesign the auction-based ad system pioneered by Google.

13:04At present, Perplexity's AI chatbot gives a comprehensive response to user questions based on information from the internet. Below this, Perplexity offers suggested follow-up queries. Under its new advertising model, brands will be able to bid for a sponsored question, which features an AI-generated answer approved by the advertiser. Perplexity has apparently been talking with companies including Nike and Marriott, and is hoping to roll out this ad system by the end of the year. And so, of course, the question is, will the AI summary model make there be less of a market for advertising and less of a pathway to publishers in general, or is it just going to be different?

13:36I'm not exactly sure, but it is a very big question that will have a much bigger impact on everyone's experience of the web than it might seem at first. Anyways, friends, that is going to do it for today's AI Daily Brief. Until next time, peace.