Tech Brew Ride Home - Mon. 09/30 – Controversial AI Bill Vetoed

Episode Date: September 30, 2024

That controversial AI bill in California has been vetoed by Governor Newsom. Is even Apple now thinking that its Vision Pro strategy might need a rethink? What really is OpenAI’s situation right now, and this time I’m talking money-wise? And the strange resurrection of the point and shoot camera. Links: California’s Gavin Newsom Vetoes Controversial AI Safety Bill (WSJ) California Passes Law Protecting Consumer Brain Data (NYTimes) Meta’s New Headsets Show Apple Has Lost Its Way With the Vision Pro (Bloomberg) OpenAI’s Complex Path to Becoming a For-Profit Company (WSJ) Songs by Adele, Bob Dylan, Green Day, Many More Blocked by YouTube in Legal Dispute (Variety) This Camera Went Viral Two Years Ago. You Still Can’t Buy One (Bloomberg) 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 Monday, September 30th, 2024. I'm Brian McCullough today. That controversial AI bill in California has been vetoed by Governor Newsom. Is even Apple now thinking that its Vision Pro strategy might need a rethink?

What really is open AI situation right now? And this time I'm talking money-wise. And the strange resurrection of the point-and-shoot camera. Here's what you missed today in the world of tech. California Governor Gavin Newsom has vetoed AI safety bill SB 1040. 47 saying it applies only to large AI models and doesn't account for if deployment is high risk. This was that AI bill that a lot of the AI community and AIVCs were very worried about, quoting the journal. California Governor Gavin Newsom has vetoed a controversial artificial intelligence safety bill that pitted some of the biggest tech companies against prominent scientists who developed the technology. The Democrat decided to reject the measure

because it applies only to the biggest and most expensive AI models and doesn't take into account Whether they are deployed in high-risk situations, he said, in his veto message. Smaller models sometimes handle critical decision-making involving sensitive data, such as electrical grids and medical records, while bigger models at times handle low-risk activities such as customer service. Had Newsom signed the bill into law, it would have laid the groundwork for how AI is regulated across the U.S. as California is home to the top companies in the industry. Proposals to regulate AI nationally have made little progress in water.

Washington. The governor announced that he is working with leading AI researchers, including Phi Phi Lee, a Stanford University professor who has worked at Google and recently launched a startup called World Labs to develop new legislation he is willing to support. Newsom signed other AI-related bills into law this year. They include one requiring AI developers to label content created by their technology and others, regulating election-related deepfakes and giving performers more control over digital replicas of themselves. The bill he vetoed SB 1047 would have required developers of large AI models to take, quote, reasonable care to ensure that their technology didn't pose and, quote, unreasonable risk of causing or materially enabling a critical harm.

It defined that harm as cyber attacks that cause at least $500 million in damages or mass casualties. Developers also would have needed to ensure their AI could be shut down by a human if it started behaving dangerously. The bill applied to AI models that met a certain computing power threshold and cost at least $100 million to train the estimated cost of OpenAI's biggest model, GPT4. Any company doing business in California, regardless of where it is headquartered, was covered. Google, Meta, Microsoft, and OpenAI raised concerns about the bill saying it imposed vague standards in the name of safety. Many smaller companies joined them, arguing the bill would chill innovation by discouraging large developers from making their models available to the public, fracturing a startup ecosystem

that relies on such openness to innovate. The bill would have prohibited developers from releasing large AI models, quote, if there is an unreasonable risk that the technology will cause or materially enable a critical harm. Tech companies said the language opened a legal minefield. Nobody knows what that means. So you're basically leaving it to a court that's non-technical, said Martin Casado, a partner overseeing AI investments at venture capital firm Andresen Horowitz, end quote. One more real quick, though, it's not quite related. Governor Newsom also signed the California Neuro Rights Act into law that treats neural data as personal sensitive information aiming to protect brain data from misuse. Quoting the Times. On Saturday,

Governor Gavin Newsom of California signed a new law that aims to protect people's brain data from being potentially misused by neurotechnology companies. A growing number of consumer technology products promised to help address cognitive issues, apps to meditate, to improve focus, and to treat mental health conditions like depression. These products monitor and record brain data, which encodes virtually everything that goes on in the mind, including thoughts, feelings, and intentions. The new law, which passed both the California State Assembly and the Senate with no voter opposition, amends the state's current personal privacy law known as the California Consumer Privacy Act by including neural data under personal sensitive information. This includes data generated by a user's

brain activity and the meshwork of nerves that extends to the rest of the body. and quote. Mark German newsletter Monday this weekend, Mark took a look at the fallout from the demo of those meta- Orion prototype smart glasses. Mark straight out came out and said, even though it was just a demo, it made even him think that Apple's strategy with Division Pro might be a mistake. Quote, Apple seems aware that it needs to rethink its approach to headsets, but there isn't consensus on how to do that, I'm told. As of now, the company's Vision Products Group is evaluating a few different options, including the status quo route. This would involve keeping the Vision Pro more or less the same, but focusing on a less expensive

version. Apple could bring down the cost with cheaper materials, lesser internal technologies, and lower-quality displays. The company would also follow up with a second-generation version of the original Higher-end Vision Pro that has a new chip and Apple Intelligence. The Smart Display Route. In this scenario, Apple would remove the onboard computer and external battery from the Vision Pro and shift many of the internal functions over to the iPhone. This would make the iPhone more valuable, reduce the weight and heat of the headset, and more importantly, eliminate several hundred dollars worth of components to bring down the price. Then there's the smart glasses route. This would mean developing a product that's closer to meta's hit collaboration with Rayban, Smart Glasses

Without AR. Apple could use its expertise in chips and audio, as well as its growing collection of AI tools to make a compelling device. This would essentially be a Me Too product, but also something akin to and AirPods on steroids, and Apple fans would probably eat it up. There's then the AI and AirPods route. Apple is working on a new version of the AirPods Pro that uses external cameras and artificial intelligence to understand the outside world and provide information to the user. This would essentially be the smart glasses path, but without actual glasses. Finally, there's the Holy Grail route. The ultimate goal is standalone augmented reality spectacles that come with high-performing lenses, a battery system onboard computer,

cameras, eye tracking, and other components built in, all while still being the size and weight of normal glasses. This has long been chief executive officer Tim Cook's dream, but Apple previously paused development of such a product because it was just too big a challenge. Apple could ultimately release some or all of these products, but it's going to need to accelerate its work and fast. If not, Apple will risk losing out on a product category that could transform the way people use technology, end quote. Speaking of Apple, the journal says that Apple is no longer in talks to participate in OpenAI's funding round that is expected to raise as much as $6.5 billion and close as soon as next week. Not sure we can or need to make anything out of that particularly right now, but it seems like that round is coming, which is making people take a hard look at where OpenAI is at the moment

financially. The New York Times has been working their sources and say OpenAI had, $300 million in monthly revenue in August up 1,700% since early 2023. They had 350 million monthly active users in June and expect around $3.7 billion in annual sales this year, but also, of course, $5 billion in losses. Reuters worked their sources to report that OpenAI expects its revenue to grow to $11.6 billion next year. ChatGPT has 10 million paying users and could bring in $2.7 billion in revenue this year. year, up from 700 million in 2020. But again, the question really is, is that growth sustainable if the cost of providing the product you provide is so expensive? So the journal took a look at the challenges

OpenAI is facing if it indeed becomes a for-profit company, which I guess it needs to do to address all that. This includes the complexity of splitting assets with the nonprofit arm, which, I hadn't thought of this might continue to exist, quoting the journal. The chat GPT maker is in the midst of raising $6.5 billion from bankers, including Microsoft and Nvidia, along with venture capital firms and a United Arab Emirates state-backed company. And a central provision of talks is that OpenAI, currently governed by a charitable non-profit, must within two years become a public benefit corporation. That means its mission is to earn profit while creating social good.

If it doesn't, investors could take back their money. To get there, it will have to deal with regulatory requirements in at least two states, determine how to award equity in the for-profit company, and split assets with the non-profit entity, which will continue to exist. This kind of transaction is incredibly complex and would involve a large number of legal and regulatory hurdles that would need to be navigated, said Karen Blackenstone, general counsel at the investment firm Hanger Management, and an attorney specializing in technology and tax-exempt organizations. Open AI is a Delaware-based corporation and would need to change its structure legally under that

state's law. Given Delaware's business-friendly legal regime, making that change would be the easiest part of the process, said Jill Horwitz, a law professor and founding faculty director of the Lowell-Milken Center for Philanthropy and Nonprofits at the University of California, Los Angeles. The more complicated part is what would happen to Open AI's assets. When such a conversion takes place, it can't simply shift assets from a nonprofit to a for-profit. The nonprofit is legally required to end up with assets, including any cash and securities, at least as valuable as those it turns over to the for-profit. In effect, OpenAI's operations would likely be sold to the for-profit company or its investors with the charity retaining the proceeds.

Assets, quote, previously donated to the public benefit cannot be repurposed to private benefit without compensating the public for the loss, said Alexander Reed, a partner at the law firm Baker Hostetler. It couldn't be determined how much the remaining nonprofit would receive in such a transaction, given that OpenAI is expected to be valued at around $150 billion after the current fundraising round, the amount could be immense. That determination is particularly complicated because the nonprofit might own some of OpenAI's patents on AI technology, which would need to be valued in the conversion. Because most of Open AI's operations are in California, that state's attorney general would have jurisdiction to

ensure its charitable assets were protected, according to legal experts. OpenAI hasn't said, what the continuing nonprofit would do once a profit-seeking company takes over all of its operations. The current nonprofit has supported research on universal basic income programs and made charitable grants focused on technology and social equality. According to its website, it could continue to fund such work in the future, using its stake in the for-profit Open AI as a source of income. One important issue open AI will have to resolve in the conversion is what rights the nonprofit will have to AI technology that the new company develops, end quote. If you've searched for some music videos from some high-profile artists recently, you might

have come away disappointed. Why? Well, quoting variety, songs by Adele, Bob Dylan, Green Day, R.E.M., Burnaboy, Rush, and many others are currently unplayable on YouTube in the U.S. due to a legal dispute between the platform and the performing rights organization CSAC. Attempts to play many but not all songs by those artists on Saturday met with the following message. This video contains content from CSAC. It is not available in your country. A similar dispute between Universal Music Group and TikTok raged on for several months earlier this year before being resolved. In a statement to variety, a YouTube rep said, We have held good faith negotiations with CSAC to renew our existing deal.

Unfortunately, despite our best efforts, we were unable to reach an equitable agreement before its expiration. We take copyright very seriously, and as a result, content represented by CSAC is no longer available on YouTube in the U.S. We are in active conversations with CSAC and are hoping to reach a new deal as soon as possible. Reps for CSAC did not immediately respond to requests for comment. A source close to the situation tells variety that the previous deal actually does not expire until next week and suggests that YouTube's move is a negotiating tactic. CESAC is far smaller than ASCAP and BMI with approximately 30,000 members and 1.5 million compositions, while ASCAP has nearly 800,000 members, but as the caliber of art.

affected by the block shows, it represents a comparatively large percentage of the marketplace, end quote. Finally today, I think we've spoken about this before, but maybe not. Back in September of 2022, the X-100 series of point-and-shoot digital cameras went viral on TikTok. This has led to huge profits for Fuji film. Imagine that, huge profits in camera hardware that isn't a smartphone, quoting Bloomberg. In September 2022, the photographer and TikTok influencer Kylie Kattich posted a video showing off her Fuji film X100V. I just found a camera that will change your life, she told, hundreds of thousands of followers at the time, as photos of her infant flashed by. With the warm, yellowish glow, the images looked like they could have been

pulled out of a photo album from a generation ago. The camera itself, with its jacket, pocket-sized aluminum body, had distinctly vintage features and came programmed with simulations that mimic the look of photographic film without any editing. It looks like a film camera. It has settings like a film camera, Kadditch said in the video, holding up the $1,400 point-and-shoot. It's digital. Within a day, her post had racked up dozens of comments. Consider me influenced, I am deaf-getting. One person replied, this is the coolest thing ever, said another. The endorsement and others like it reverberated through TikTok's sub-communities, auto enthusiasts, brought their X-100 Vs to sports car meetups. Musicians took theirs around the world to document the touring lifestyle. Tech geeks, travel

junkies, and fashion influencers all wanted an X100V, and once they got one, they let their own followers know. The effect on the company's inventory, already strained by a global chip shortage, was drastic. By November 2022, Fujifil stopped accepting online orders for the camera, citing its inability to keep up with demand. Buyers turned to resale platforms such as eBay, where X100Vs were selling for twice their retail value. Two years later, TikTok, The Fedox's obsession hasn't slowed, and the manufacturer is still reaping the benefits. Fujifilm's photo imaging segment brought in $3.3 billion for the fiscal year that ended in March, a 14 and a half percent year-over-year increase, and was the company's most profitable division for the first time since at least 2006,

when Bloomberg started tracking earnings by segment. Those sales were driven in part by the X100V's successor, sporting and upgraded sensor and other new features. The X100 V1 or VI sold out almost a massive. immediately upon its release in February. Both generations remain unavailable on Fujifilm's website. The company is embracing the camera's coveted status, and a May earnings call, Chief Executive Officer Tieti Gotu said the X100 VI's inventory situation was normal and compared the product's resale value to that of a prestigious competitor. Lyca, a well-known German manufacturer, still maintains a very high value for both their old cameras and the cameras they sell now,

and this is our goal, Go-to said in the call. Fuji film, which declined request for interviews, said in a statement that it's accelerating production to meet demand. Our goal is that every customer who ordered an X100 VI gets it, the company said, end quote. Nothing for you today. Talk to you tomorrow.