Tech Brew Ride Home - Wed. 07/03 – Walking Back AI Products

Episode Date: July 3, 2024

The deploying and then walking back of AI products and features is becoming something of a routine at this point. Apple is joining OpenAI’s board, kinda-sorta. Meta outlines 3D Gen. Proof that VC funding is coming back, baby. And, of course, the Weekend Longreads Suggestions. Sponsors: A Better Paradise Podcast Links: Figma Disables AI App Design Tool After It Copied Apple’s Weather App (404 Media) Figma pulls AI tool after criticism that it ripped off Apple’s design (The Verge) Apple Poised to Get OpenAI Board Observer Role as Part of AI Pact (Bloomberg) Meta drops ‘3D Gen’ bomb: AI-powered 3D asset creation at lightning speed (VentureBeat) Investors Pour $27.1 Billion Into A.I. Start-Ups, Defying a Downturn (NYTimes) Weekend Longreads Suggestions Internet Browsers Are Getting a Makeover for the Workplace (WSJ) A.I. Begins Ushering In an Age of Killer Robots (NYTimes) Learn more about your ad choices. Visit megaphone.fm/adchoices

On April 4th, 2023, around 2 in the morning, a man was found stabbed multiple times on a sidewalk in downtown San Francisco. Hey, who did this to you? What happened next turned the story into a political firestorm. Reports have identified the victim as Bob Lee, the founder of Cash App. From Bloomberg Podcasts, this is Foundering, the Killing of Bob Lee, beginning April 16. Welcome to the Tech meme right home for Wednesday, July 3rd, 2024. I'm Brian McCullough today. The deploying and then walking back of AI products and features is becoming something of a routine at this point. Apple is joining OpenAI's board, kind of sort of. Meta outlines 3D Gen. Proof that VC funding is coming back, baby. And of course, the weekend long read suggestions. Here's what you miss today from the world of tech. Well, this is becoming something of a trend. Remember how Microsoft had to put. pump the brakes on or delay or even walk back some of their AI features recently. I think I told you recently that Figma had launched an AI app design tool called MakeDesign.

I hope I did, because Figma has already suddenly disabled Make Design. After a user showed it copied Apple's Weather app when it was asked to design a weather app. Quoting TechCrunch. Figma CEO Dylan Field says the company, company will temporarily disable its make-design AI feature that was said to be ripping off the designs of Apple's own weather app. The problem was first spotted by Andy Allen, the founder of Not Boring Software, which makes a suite of apps that includes a popular skinnable weather app and other utilities. He found by testing Figma's tool that it would repeatedly reproduce Apple's

weather app when used as a design aid. Alan had taken to X, formerly Twitter, to accuse Figma of heavily training its tool on existing apps, and accusation field now denies. The make design feature is available within Figma's software and will generate UI or user interface layouts and components from text prompts. Just describe what you need, and the feature will provide you with a first draft is how the company explained it when the feature launched. The idea was that developers could use the feature to help get their ideas down quickly to begin exploring different design directions and then arrive at a solution faster, Figma said.

The feature was introduced at Figma's Config conference last week where the company explained that it was not trained on Figma content, community files, or app designs. Field notes in his response on X. In other words, the accusations around data training in this tweet are false, he said. But in its haste to launch new AI features to remain competitive, the quality assurance work that should accompany new additions seems to have been overlooked. Allen's discovery that Figma essentially seemed to be copying other apps led to increased concern among the design community.

Just a heads up to any designers using the new Make Designs feature that you may want to thoroughly check existing apps or modify the results heavily so that you don't unknowingly land yourself in legal trouble, Alan warned others on X. Field responded by clarifying that Make Design uses off-the-shelf large language models combined with, quote, systems we commissioned to be used by those models. He said the problem with this approach is that the variability is too low. Within hours of seeing Alan's tweet, we identified the issue, which was related to the underlying design systems that we are, created. Field wrote on X. Ultimately, it is my fault for not insisting on a better QA process for this work and pushing our team hard to hit a deadline for config, end quote. Apple was not immediately available for comment. Figma pointed to Fields' tweets as its statement on the matter. Field says Figma will temporarily disable the make design feature until the team is confident

it can, quote, stand behind its output. The feature will be disabled as of Tuesday and will not be re-enabled until Figma has completed a full QA pass on the feature's underlying design. design system, end quote. Meanwhile, Figma's CTO Chris Rasmussen told this to the verge, quote, we did no training as part of the generative AI features, Rasmussen said. The features are, quote, powered by off-the-shelf models and a bespoke design system that we commissioned, which appears to be the underlying issue, end quote. That generally matches something he said on Monday on X in response to a user who suggested make designs was trained on existing apps. As we shared when we launched Figma AI last week, there was no training as part of this feature,

or any of our generative features, he wrote. We are looking into what extent the similarities are a function of the third-party models we are using versus the design system we commissioned to be used by the models, and we will address as needed, end quote. So that makes me wonder what would be worse? What would be the worst implication if it actually bared out here? Would it be worse that Figma eventually was proven to be copying other folks' designs with their AI, or that any of the off-the-shelf AI is going to copy other folks' designs, because then you can't take things off-the-shelf with any degree of confidence, right? This is interesting. Mark German says Apple will get an observer role on OpenAI's board of directors,

with Phil Schiller chosen for the position as part of the partnership that OpenAI announced with Apple last month. Quote, as a board observer, Schiller won't be serving as a full-fledged director, said the people who not to be identified because the matter isn't public. The board observer role will put Apple on par with Microsoft, OpenAI's biggest backer and its main AI technology provider. The job allows someone to attend board meetings without being able to vote or exercise other director powers. Observers, however, do gain insights into how decisions are made at the company. Having Microsoft and Apple sit in on board meetings could create complications for the tech giants, which have been rivals and partners over the decades.

Some Open AI board meetings will likely discuss. future AI initiatives between OpenAI and Microsoft, deliberations that the latter company may want Schiller excluded from. Board observers often do oblige and exit meetings during discussions that are seen as sensitive. Though Schiller doesn't lead Apple's artificial intelligence initiatives, his veteran role as a steward of the company's brand makes him well suited to the observer job, the 64-year-old stepped down as head of marketing in 2020, and his current title is Apple Fellow. In addition to overseeing the app store, he still manages the company's product launch events, outside of Apple. He is a director at medical technology company, Illumina Inc., end quote.

Meta has shared its research on meta 3D gen, a system that creates high-quality 3D assets from text descriptions in less than a minute. So not just images, but also not movies or videos, 3D assets, quoting Venture Beat. This development represents a significant advance in generative AI for 3D graphics, with the potential to transform industries from video game development to industrial design and architecture. The new system combines two key components, meta 3D asset gen, which creates 3D meshes, and meta3D texture gen, which generates textures. Working in concert, these technologies produce 3D assets with high-resolution textures and physically based rendering PBR materials. According to meta, the process is 3 to 10 times faster

than existing solutions. 3DGen offers 3D asset creation with high-prompt fidelity and high-quality 3D images and textures in under a minute. The researchers explained in their paper, this speed and quality could streamline workflows across multiple industries, reducing production time and costs while enhancing creative possibilities. One of meta3D gen's most significant features is its support for PBR material, allowing for realistic relighting of generated 3D objects. This capability is crucial for integrating AI-generated assets into real-world applications and existing 3D pipelines. The integration of PBR materials represent a significant advancement in AI-generated 3D content,

traditional AI-generated 3-D assets often struggle with realistic lighting and material properties, limiting their usefulness in professional workflows. By incorporating PBR, meta-3-Gen allows generated assets to react realistically to different lighting conditions, making them far more versatile and applicable in various contexts from video game environments to architectural visualizations. This breakthrough could potentially bridge the long-standing gap between AI-generated content and professional 3D workflows. It enables seamless integration of AI-created assets into existing pipelines, potentially accelerating the creation of virtual environments and digital twins across

various industries. The technology also has significant implications for the growing fields of augmented and virtual reality. By lowering the barriers to creating high-quality 3-D assets, meta-3-D gen could accelerate the development of immersive experiences for education, training, and entertainment. However, the advent of this technology also raises questions about the future of 3D modeling as a profession. While meta3D gen is likely to enhance the productivity of 3D artists, it may also reduce the demand for certain types of routine 3D asset creation. This shift could push professionals to focus more on creative directions and complex custom work. That AI still struggles to replicate, end quote. So I'm on the boat now, the SS Badger Car Ferry going across

Lake Michigan. I assume we're using Starlink for the internet here. And if so, I have to say the speed thus far is very impressive. Let me stick the mic out the window here and see if you can hear it. That is not any sound effects sweetening. I am literally in a cabin stuck the mic out the porthole. Anywho, confirmation of what you might have suspected when it comes to VC investing. We're sort of back, and AI is leading the way. According to Pitchbook, U.S. startups raised $56 billion in Q2 of 2024,

which is up 57% year over year and the highest three-month haul in two years. AI startups raised $27.1 billion, which is nearly half of the total. Quoting the New York Times, AI companies are attracting huge rounds of fundraising reminiscent of 2021, when low interest rates pushed investors away from taking risks on tech investments. In May, Corweave, a provider of cloud computing services for AI companies raised $1.1 billion, followed by $7.5 billion in debt, valuing it at $19 billion. Scale AI, a provider of data for AI companies, raised $1 billion, valuing it at $13.8 billion,

and XAI founded by Elon Musk, raised $6 billion, valuing it at $24 billion. Such financing rounds have boosted the industry's overall deal-making by dollar amount and number of deals, said Kyle Sanford, a research analyst at Pitchbook. It's not declining anymore, he said. the bottom has already fallen out. The activity has prompted some venture capital investors to change their message. Last year, Tom Leverro, an investor at IVP, predicted a, quote, mass extinction event for startups and encouraged them to cut costs. Last week, he declared that era over and christened this time the Great Awakening, encouraging companies to, quote, poor gas on growth,

particularly around artificial intelligence. The AI train is leaving the station and you need to be on it, he wrote on X, end quote. give you another anecdote here, which you should take with a grain of salt, of course, but on the TechMeme back end, we see all of the new funding round announcements. I only occasionally tell you about the most interesting raises, but TechMeme, the website, posts almost all of them. And the anecdote is that the TechMeme editors have been commenting about the absolute flood of funding news lately. It's almost starting to feel like the boom times of three years ago once again. From my perspective, I think what we're seeing is a lot of the seed or stealth AI companies that got started

over the last year are coming out or raising A rounds. A lot of the AI companies that were founded before Chat GPT are raising B and C and D rounds. And most encouragingly, lots of companies, even non-AI ones that were sitting on A's and Bs that couldn't raise their next rounds, are starting to be able to do just that. Time for the weekend long read suggestions. First up, Speaking of fundraising, one of the weirdly hottest spaces recently for startups has been web browsers, and not just the AI-flavored arc browser either. There are a lot of enterprise-grade browser startups out there all of the sudden. Why?

Well, quoting the Wall Street Journal. While adoption is low today, research firm Gartner estimates that by 2030, enterprise browsers will become the core platform for delivering workforce productivity and security software on devices. Pfizer said it has deployed the Island Enterprise browser in some use cases, and it is aiming for a more scaled-out deployment in the future. Coleman said he was skeptical at first about paying for a browser, since traditional browsers are available at no extra cost, but he has found using Island to be more cost-effective since he is able to consolidate some of the other security measures in the tech stack. It's not going to solve everything, but when you're looking at improving security, saving money,

looking for mature capabilities beyond the basic browser, I think these will be very important in the future, Coleman said. One of the benefits, Coleman said, is the ability to easily control what data employees can access within apps based on their role in the organization, something that is possible in commercial browsers, but only with a heavy technical lift. Island's chief customer officer Brandon Rogers said businesses can also control how users interact with data, for example, whether they are allowed to copy and paste it into other apps or upload it to chat GPT, end quote. And finally today, it was the Crimean War and the U.S. Civil War that gave the world hints

that technology had changed warfare forever, hence that nobody really heated, and thus we got the strategy of World War I. It was also the Spanish Civil War that was the dress rehearsal for the technology of aerial and mobile warfare that came to bear in World War II. Well, from the New York Times, how Ukraine is using air. AI, code found online, and hobbyist computers to weaponize consumer tech and build low-cost weapons like autonomous drones and machine guns. So, if the future of warfare is some sort of Terminator-style robot battlefield, Ukraine is showing the world that the future is now. Quote, Virey is just one of many Ukrainian companies working on a major leap forward in the weaponization

of consumer technology driven by the war with Russia. The pressure to outthink the enemy, along with huge flows of investment, donations, and government contracts has turned Ukraine into a Silicon Valley for autonomous drones and other weaponry. What the companies are creating is technology that makes human judgment about targeting and firing increasingly tangential. The widespread availability of off-the-shelf devices, easy-to-design software, powerful automation algorithms, and specialized artificial intelligence microchips has pushed a deadly innovation race into uncharted territory, fueling a potentially new era of killer robots. The most advanced versions of the technology that allows drones and other machines to act

autonomously have been made possible by deep learning, a form of AI that uses large amounts of data to identify patterns and make decisions. Deep learning has helped generate popular large language models like OpenAIs GPD4, but it also helps make models interpret and respond in real-time to video and camera footage. That means software that once helped a drone follow a snowboarder down a mountain, can now become a deadly tool. In more than a dozen interviews with Ukrainian entrepreneurs, engineers, and military units, a picture emerged of a near future,

when swarms of self-guided drones can coordinate attacks and machine guns, with computer vision, can automatically shoot down soldiers. More outlandish creations like a hovering unmanned copter that wields machine guns are also being developed. The weapons are cruder than the slick stuff in science fiction blockbusters, like The Terminator and its T-1000 Liquid Metal Assassin, but they are a step toward such a future. While these weapons aren't as advanced as expensive military-grade systems made by the United States, China, and Russia, what makes the development significant is their low cost, just thousands of dollars or less, and ready availability.

Except for the munitions, many of these weapons are built with code found online and components such as hobbyist computers like Raspberry Pi that can be bought from Best Buy and a hardware store. Some U.S. officials said they worried that the abilities could soon be used to carry out terrorist attacks, end quote. Here's the deal for the rest of the week. I am taking the next two days off because of July 4th, at least taking the days off from doing fresh shows. I will leave you with my two Internet History podcast episodes about Bill Von Meister. I know I shared those with you a couple years ago around Thanksgiving, but they remain the most produced narrative audio I've ever done. So if you've never listened to the story of Bill von Meister, please do. Otherwise, I will talk to you on Monday. My fellow Americans, enjoy your fourth.