Should AI Gains Mean a 32-Hour Workweek?

0:00Today on the AI Breakdown, Apple has made another AI-related acquisition. Before that, on The Brief, Bernie Sanders has proposed a 32-hour workweek based on productivity gains from AI. The AI Breakdown is a daily podcast and video about the most important news and discussions in AI. Go to Breakdown.network for more information about our YouTube channel, our newsletter, and our Discord.

0:24Welcome back to the AI Breakdown Brief, all the AI headline news you need in around five minutes. Quick disclosure here, we haven't had this come up very much, but I want to explain my perspective on covering politics on this show. What you will not hear on this show is me taking a lot of positions on policy unless they are perhaps directly about AI. But what you will hear is coverage when AI becomes part of the political discourse. I am interested constantly in how a new technological and sociological phenomenon like AI finds its way into left, right, and independent politics. And that's what has me interested in this week's announcement from Senator Bernie Sanders.

0:58I think most of you will know who Bernie Sanders is at this point, but he is a former U.S. presidential candidate who came very close to winning the nomination from Hillary Clinton, and who is a standard-bearer for the far-left progressive movement. Bernie tweeted a couple days ago,

1:23The announcement reads, Senator Bernie Sanders, chairman of the Senate Committee on Health, Education, Labor, and Pensions, today announced that he will introduce legislation to establish a standard 32-hour workweek in America with no loss in pay, an important step towards ensuring that workers share in the massive increase in productivity driven by artificial intelligence, automation, and new technology. Senator Sanders said, Today, American workers are over 400 % more productive than they were in the 1940s, and yet millions of Americans are working longer hours for lower wages than they were decades ago.

1:50That has got to change. The financial gains from the major advancements in artificial intelligence, automation, and new technology must benefit the working class, not just corporate CEOs and wealthy stockholders on Wall Street. It's time to reduce the stress level in our country and allow Americans to enjoy a better quality of life. It is time for a 32-hour workweek with no loss in pay. CNBC reflects that, quote, the bill comes after months of business leaders such as JPMorgan Chase CEO Jamie Dimon and Microsoft co-founder Bill Gates predicting that people could within decades work as little as three days a week due to innovations in AI and automation.

2:21Republicans, unsurprisingly given that this plan comes from Bernie Sanders, are not particularly into it. However, their reasons don't seem to be about a disagreement about AI productivity or who should benefit from AI productivity, but instead because of the cost it potentially imposes on small businesses and industries which require different types of working structures such as retail stores, which need to be open longer. And indeed, this drawing the line between different types of businesses was a really common theme. For example, Senator Bill Cassidy of Louisiana shared skepticism that AI's advances were going to be equally distributed.

2:50Said Cassidy, a mom-and-pop restaurant is not really seeing increased productivity from AI. They're having trouble finding enough people to fill shifts. So here's what I think is fascinating about this conversation. It is very, very different than the other political conversations that we're having so far about artificial intelligence. Those conversations tend largely to be about regulation and the relationship between big tech and the government. In fact, many of them echo longer-term themes of big tech and its power that we've been discussing for a decade or more now. There is also, of course, some amount of ex-risk conversation and concerns about AI safety.

3:22What Bernie Sanders, however, is bringing up here, and instigating a conversation around, is how the benefits of AI are distributed. It carries with it an implicit assumption that AI is here to stay, that it is having positive impact on productivity, and that because of that, we should make different decisions as a society about the social contract. Now, a lot of the counterpoints being made are refreshingly substantive relative to US politics. How does this impact something like retail? What about small business owners that really are having a hard time filling shifts? Those are, of course, the political negotiations and practicalities that would need to be worked out.

3:56But I feel strongly that whatever answers we come to as a society, it is our prerogative in the context of this new technology to ask whether it changes what we believe the appropriate social contract is. Whether we think that because of AI, giving a 32-hour workweek should entitle you to the same gains that 40 hours a week once did. I like very much when we save our breath for good conversations to have. And I tend to think, almost no matter what perspective you're coming from on the answer to this one, that it firmly falls in that good conversation to have territory. Another conversation that some are having is whether celebrities should be resurrected by AI or not.

4:32Although frankly, I tend to think that this is one where the ship is sailed, and it's going to be very, very hard to get this toothpaste back in the tube. We have later this year an Elvis experience coming to the stage, and at South by Southwest this year, a company called Authentic Brands Groups, which had acquired Marilyn Monroe's intellectual property in 2011, created something that they call Digital Marilyn, which uses AI to recreate her voice and affectations in real time. Right now, these mostly function as a Rorschach test for what people think about AI, but whatever you think, I would expect more, not less of these, especially in the context of artists increasingly cashing out and selling their music catalogs and other types of IP rights while they're still alive, leading to a lot more legal openness for this type of avatar duplication.

5:16Over in the world of Wall Street, AI continues to eat finance as it is eating everything else, the latest example being that Morgan Stanley has named a head of artificial intelligence for the first time. The company is promoting Jeff McMillan, who has worked in their wealth management division for some years, to oversee AI across the firm, including all of their internal implementations. Last year, McMillan had overseen a project to create a Morgan Stanley customized solution for employees based on GPT-4. The announcement reads, In his new role, Jeff will coordinate across the firm to ensure we have the appropriate AI strategy and governance in place.

5:49In doing so, he will partner with the business units and infrastructure areas to identify and prioritize AI opportunities, help position the firm within the flow of AI development across the industry, and ensure that Morgan Stanley continues to be a well-respected innovator in AI. Finally today, over in startup land, confirmation of a deal that the information first reported on back in February, Together AI, which is a company that helps AI developers access NVIDIA chips, particularly enterprise customers, to train and fine-tune their models, has raised a$106 million round, valuing the company at$1.25 billion.

6:20The round was led by Salesforce. That is going to do it for today's AI Breakdown Brief. Next up, the main AI breakdown. Welcome back to the AI breakdown. down, Apple has been a laggard in the AI arms race that has engulfed big tech over the last year, year and a half. But that seems to all be changing. Today, we're going to talk about a new acquisition from the company first reported by Bloomberg, and then we're going to check in to see what the state of what we know about Apple's AI efforts is. First of all, the company that Apple has acquired is called Darwin AI. And this purchase was made earlier this year before any of the big changes which came just a couple of weeks ago.

6:59Bloomberg characterizes Darwin AI as having developed AI technology for visually inspecting components during the manufacturing process. It has customers across a range of industries. However, what potentially caught Apple's eye might be a technology that they've developed to make AI systems both smaller and faster. An Intel blog from back in 2020 describes some of what they were working on. They write, As a student pursuing a doctorate in systems design engineering at the University of Waterloo, Alexander Wong didn't have enough money for the hardware he needed to run his experiments in computer vision.

7:29So he invented a technique to make neural network models smaller and faster. Said Darwin AI's CEO Sheldon Fernandez, We build up a very sophisticated understanding of a neural network, and then we use AI a second time to generate a new family of neural networks that's as good as the original, a lot smaller, and can be explained. One of the examples they gave, an automotive customer of Darwin AI, was troubleshooting an automated vehicle with a strange tendency to turn left when the sky was a particular shade of purple. Darwin AI's solution, which it calls generative synthesis, helped the team recognized how the vehicle's behavior was affected by training for certain turning scenarios that had been conducted in the Nevada desert, coincidentally when the sky was that purple hue.

8:03Now when you take Darwin in that light, all of a sudden this acquisition starts to make a lot more sense. A big part of the presumption last year when Apple wasn't diving into AI was that part of the reason might be the company's preference for running software on device rather than relying on the cloud. Companies like Google, Amazon, and Microsoft are of course all competing for cloud business, but for both privacy and performance reasons, Apple likes to do things on your mobile phone, in your laptop. There has been a steady push, not just from Apple, but from others as well, to make AI models that can actually run on device without relying on the cloud, and to the extent that Darwin AI helps that process, you can see how they would become a valuable acquisition.

8:41Alexander Wong, who's the AI researcher from the University of Waterloo that we were just discussing, joined Apple as a director in the AI group as part of the deal. When asked about the deal, an Apple representative said only that it, quote, by smaller technology companies from time to time. So that is the new news, but let's talk about what has been happening. One of the biggest updates we got was from two weeks ago, when Apple pulled the plug on its electric car project. On the one hand, this represented a ton of sunk costs for Apple. The company is estimated to have spent more than$10 billion over the course of 10 years on this initiative.

9:11On the flip side, there have long been many skeptics of this particular work, and so even beyond the fact that it seemed like they were rededicating resources to AI, Wall Street was excited for them to not have their focus over on cars. But it was indeed AI that seems to have been the biggest beneficiary of the switch, with many of the 2 ,000-person team working on Project Titan shifting over into those generative AI efforts. Pretty soon after this, Apple CEO Tim Cook also started talking about AI a lot more. The way that Petapixel described it on March 4th was that in the blink of an eye, Apple is all in on AI.

9:41They write, In just a month, Apple went from basically only ever mentioning AI in passing to shifting into high gear as CEO Tim Cook spoke twice about the company's endeavors there within three weeks. In early February, Cook said, Our MO has always been to do the work and then talk about the work and knock it out in front of ourselves. And so we're going to hold to that as well. But we have got some things we're incredibly excited about that we'll be talking about later this year. However, soon after, in a shareholder meeting, Cook said that the company was planning on, quote, breaking new ground in generative AI this year, calling generative AI, quote, another technology we believe can redefine the future.

10:12Cook said, we believe it will unlock transformative opportunities for our users when it comes to productivity, problem solving, and more. But it's not just statements from Cook. When the company released its new MacBook Air, which is powered by the M3 chip, they dedicated a chunk of the press release specifically to artificial intelligence. Apple said in that release, with the transition to Apple Silicon, every Mac is a great platform for AI. Leveraging incredible AI performance, macOS delivers intelligent features that enhance productivity and creativity, so users can enable powerful camera features, real-time speech-to-text, translation, text predictions, visual understanding, accessibility features, and much more.

10:45They pointed to apps like Copilot, Canva, Adobe Firefly, CapCut, GoodNotes, and Pixelmator as apps that can use AI in the MacBook Air, and concluded with the argument that the M3 chip makes the MacBook Air, quote, the world's best consumer laptop for AI. After this announcement, Fast Company wrote a piece called Apple Just Hinted at Its AI Plans and They Could Be Game-Changing. They make the same point that Apple has significantly shifted its tune, given that it's now marketing against AI, and speculates on what AI features might be coming. The first they discuss is the one that most people are anticipating and desperately waiting for, which is an update to Siri.

11:18They write, When it comes to generative AI, the most obvious use case and the place where it's desperately needed is in Apple's voice assistant Siri. Once the pacesetter for natural language processing, Siri is now more than a decade old, and in terms of its capabilities and comprehension, it still pales in comparison to other voice assistants that came after it. Recently, a machine learning researcher at Apple was tweeting about new smaller model performance, leading tech evangelist Robert Scoble to retweet that person, saying, How do you hype people up for future Siri without mentioning Siri?

11:43Dag Kitloss, the co-founder and CEO of Siri, responded,

11:56So again, we have notoriously tight-lipped Apple starting to actually hype things up in the context of a Twitter conversation. Other ideas about what might be coming include, as Fast Company writes, System-wide textual generative AI could be used to summarize documents, emails, and long text messages. They point to Apple Photos and iMovie that could be updated for AI. But the big one, they say, is the idea of a chatbot that lives on the device, not in the cloud. They write, If Apple's publicly available paper is any indication, the company may be further along than its competitors when it comes to finding ways for our everyday devices to run these complex AI tools locally.

12:28Apple's paper states, We believe as LLMs continue to grow in size and complexity, approaches like this work will be essential for harnessing their full potential. They conclude,

12:54Right now, we are still mostly in a mode of guessing. What's clear is that Apple has shifted its focus, its attention, and, frankly, that they have the resources to compete here, even if they're a little bit slower out of the gate. Certainly, I will be listening to every rumor and interesting new bit of news that comes up, and we'll share them with you here as they become available. For now, however, that is going to do it for the AI Breakdown. Until next time, peace.