The AI Productivity Boom Finally Shows Up

Today on the AI Daily Brief, AI productivity gains appear to be finally visible in the macroeconomic data, while before that in the headlines, it's Anthropic versus The Department of War. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. Alright friends, quick announcements before we dive in. First of all, thank you to today's sponsors, Robots and Pencils, Scrunch and Blitzy. To get an ad free version of the show, go to patreon.com aidaily brief or you can subscribe on Apple Podcasts. To learn about sponsoring the show, send us a note at sponsorsideailybrief AI aidailybrief AI is also going to be your hub for all of the various projects in the AI DB ecosystem. Yesterday we announced Claw Camp, the follow up to our AIDB New Year's Resolution program that will help you learn how to build agents with OpenClaw. It looks like we'll have more than a thousand folks join in the first 24 hours, so that is very exciting to see. You can also find out about our open role for a Clarkitect. This is basically a role that is going to help me keep track of all the developments in the OpenClaw ecosystem and then test them out on real AI DB projects. And then the one other OpenClaw related project that I wanted to share with you guys today is Enterprise Claw. It is exactly what it sounds like, a way to help businesses who are thinking about the implications of openclaw and this agentic shift more broadly figure out what challenges they're going to face as they incorporate them into their businesses. This is going to be an executive sprint, basically a capacity building program for executives that's going to kick off in March and has just 20 spots available. If you go to EnterpriseClaw AI, you can sign up to join the waitlist and get more information when it comes. But with that out of the way, let's get into the show. We kick off today with a story that is getting more and more attention. Anthropic schism with the Pentagon continues to escalate as the Department of War threatens to blacklist them from the entire military supply chain. Reports that Anthropic was pushing back on terms of use for their technology began circulating last week, but boiled over in new reporting this weekend. The Wall Street Journal reported on Saturday that Claude had been used as part of the Maduro raid. The reporting didn't go into specifics as the operation is classified, but there's a suggestion that the military breached Anthropic's Terms of Use Anthropic has drawn a hard line on prohibiting Claude from being used to facilitate violence, develop weapons or conduct surveillance. Axios reported that an Anthropic executive had reached out to counterparts at Palantir, which serves Claude to the military, in order to get answers on exactly how it was used. A senior administration official told Axios that the Palantir executive informed the Pentagon because it was raised in such a way to imply that they might disapprove of their software being used because obviously there was kinetic fire during that raid. People were shot. Anthropic, for their part, denied the reporting, stating that they had not discussed the use of Claude in any specific operations. The Department of War, however, issued a very pointed statement with chief Pentagon spokesperson Sean Parnell commenting, the Department of War's relationship with Anthropic is being reviewed. Our nation requires that our partners be willing to help our warfighters win in any fight. Ultimately, this is about our troops and the safety of the American people. Further reporting from Fox News on Monday said that Secretary of War Pete Hegseth is, in their words, close to cutting ties with Anthropic and banning contractors from using their technology. A senior Department of War official said given Anthropic's behavior, many senior officials in the Dow are starting to view them as a supply chain risk. We may require that all our vendors and contractors certify that they don't use any Anthropic models. Quotes provided to Axios include a direct threat with a senior official stating it will be enormous pain in the ass to disentangle and we are going to make sure they pay a price for forcing our hand like this. Regarding the designation as a supply chain risk, Axios noted, that kind of penalty is usually reserved for foreign adversaries. Maybe the clearest comparison would be the ban on Huawei networking equipment in 2019. The reporting also reinforced that the Pentagon is dictating terms that allow AI use for all lawful purposes rather than allowing the tech companies to set their own limits. OpenAI, Google and XAI have also contracted with the Pentagon, but only one of these firms have agreed to remove restrictions on classified use. On the record, a senior official said they were confident the other three companies would agree to the all lawful use standard. However, Axio sources said the discussions are still very much undecided. Now this is a pretty hot button topic and one where a lot of folks are wading into the conversation. It may be that we come back to it at some point because if it was just a bun fight between Pete Hegseth and Dario Amadei. Who cares? What it actually is is a proxy skirmish for who gets to dictate the terms of AI's use. Is it the companies that built it, the governments in which those companies operate, or some weird combination that's constantly being negotiated? Obviously the implications of that question is pretty significant. Next up, another Chinese model slips in ahead of the anticipated Deep seq4 release. Alibaba has launched Qin 3.5, specifically the larger Qin 3.5 version. The model has a total of 397 billion parameters arranged in a mixture of experts, architecture and supports a million token context window. Alibaba said that strong benchmark results across reasoning, coding, agentix and multimodal understanding will quote empower developers and enterprises to achieve significantly greater productivity. Productivity, by the way, is the topic of our main episode, so stick around for that. The suite of benchmarks described a model that is a big improvement over Quin 3 Max thinking, but broadly in line with GPT 5.2, Opus 4.5 and Gemini 3 Pro. Alibaba didn't compare Quinn 3.5 to the latest US models, where it falls short of the state of the art. Still, the big new capability is native multimodal reasoning, allowing Quen to join Kimik 2.5 as the only open source models with this capability. Quinn 3.5 comes pretty close to Gemini 3 Pro on multimodal benchmarks, suggesting Alibaba has basically caught up to the west in this regard. What's more, Alibaba Cloud is serving the model at 1.$2 per million input tokens and 7.$2 per million output tokens, making this even cheaper than Kimik 2.5. This one might not be as big a deal on the face of it, but it continues to reinforce just how much intelligence is available for ever diminishing prices from the Chinese labs. Next up, one story that started last week, Hollywood is in full on freakout mode over the Seed Dance 2.0 videos. When the new video model from TikTok parent company ByteDance was released last week, it was pretty immediately obvious that it would shake up. Video Production director Ruari Robinson showed what the new model could do, posting a 15 second clip of Tom Cruise fighting Brad Pitt in a post apocalyptic cityscape. Deadpool scriptwriter Rhett Reese took it far more seriously, commenting that the video sent a cold shiver up his spine. He told the New York Times, for all of us who work in the industry and devoted our careers and lives to it, I just think it's nothing short of terrifying. I could just see it costing jobs all over the place. Now, a huge part of the realization here is that Chinese AI firms are not going to play ball on copyright. American AI companies have so far been fairly responsive in the face of lawsuits, modifying their models to prevent real actors and copyrighted characters from being depicted. Not so in China, where it seems the labs have absolutely zero concern for copyright, something which frankly was entirely predictable. So then, rather than lawsuits, Hollywood is trying their luck with public denouncements. The Motion Picture association issued a statement last week commenting, in a single day, the Chinese AI service Seed Dance 2.0 has engaged in unauthorized use of US copyrighted works on a massive scale. By launching a service that operates without meaningful safeguards against infringement, ByteDance is disregarding well established copyright law that protects the rights of creators and underpins millions of American jobs. ByteDance should immediately cease its infringing activity. The Screen Actors Guild quickly joined in, adding that the use of likenesses of their members is, quote, unacceptable and undercuts the ability of human talent to earn a livelihood. Seed Dance 2.0 disregards law, ethics, industry standards and basic principles of consent. Responsible AI development demands responsibility and that is non existent here. Interestingly, ByteDance did in fact acknowledge the criticism, issuing a statement to CNBC which reads, bytedance respects intellectual property rights and we have heard the concerns regarding seed dance 2.0. We are taking steps to strengthen current safeguards as we work to prevent the unauthorized use of intellectual property and likenesses by users. Now of course we have no idea what these new safeguards will look like, and even if ByteDance does something, they have a huge stake in the US market through their minority ownership of TikTok. Whether smaller Chinese labs play by those rules seems extremely skeptical to me. Now I think there is a growing sense that even with that, Hollywood and the entertainment industry as a whole is just going to have to adapt. That doesn't mean that they should give up the fight around copyright, but the idea that that's going to come with putting this particular toothpaste back in the tube is something that I think hopefully people are starting to move on from. Rhett Reese again actually tweeted about this saying Hollywood has long been a gatekeeper that keeps young and poor people away from creative levers. When a young person with no capital sets out to impress Hollywood, they will use tools like these and young Chris Nolan's will be among them and amazing stuff will result. I actually think frankly that the entertainment industry has a pretty unique ability to be a leader when it comes to figuring out how to integrate a totally and exciting mode of production while not just being totally bowled over. Lastly, today Apple has announced a mysterious event across multiple locations. Members of the press have been invited to attend on March 4th across New York, Shanghai and London. The invitation itself said nothing beyond this being a special Apple experience. Now, the event is a little unusual for Apple, who typically arrange their major releases around their worldwide developer conference in June. However, multiple sources are discussing a big lineup of hardware that could be released at the event. Bloomberg's Mark Gurman pointed to products including new MacBook Pros, MacBook Airs, a low cost, colorful MacBook, new iPad models, and the cheaper iPhone 17e that are all slated for release in the first half of this year. He suggested the idea of an in person experience implies a more low key showcase than typical Apple events. The new MacBook Pros and Airs will likely include the new M5 chip, which was unveiled in October and featured a 30% boost to unified memory bandwidth over the M4. Now when it comes to our AI folks, the thing that they will be watching for is of course whether the lineup ends up in a surprise M5 Mac mini. If so, I'm sure there will be tons of open claws that will be shopping for an upgrade. The other big question is whether we'll get the unveiling of AI Siri officially. CEO Tim Cook has said that the new Siri will be released this year. However, most believe the release will be tied to the next iOS update expected at WWDC in June. Behind the scenes, Gurman reported last week that the project is running into roadblocks. According to his sources, recent testing has revealed issues with properly processing queries and speed to respond. Gurman did note that March of this year was the original release target and that the situation remains fluid. So who knows, the event could still feature a Siri teaser or even the full release. If problems are resolved by then. We'll keep an eye on that. But for now, that is going to do it. For today's headlines, Next up, the main episode Most companies don't struggle with ideas. They struggle with turning them into real AI systems that deliver value. Robots and Pencils is a company built to close that gap. They design and deliver intelligent cloud native systems powered by generative and agentic AI with focus, speed and clear outcomes. Robots and Pencils works in small high impact pods. Engineers, strategists, designers and applied AI specialists working together to move from idea to production without unnecessary friction. Powered by Roboworks, their agentic acceleration platform teams deliver meaningful results, including initial launches in as little as 45 days, depending on scope. If your organization is ready to move faster, reduce complexity, and turn AI ambition into real results, Robots and Pencils is built for that moment. Start the conversation@rootsandpencils.com aidaily brief that's robotsandpencils.com aidDaily Brief Robots and Impact at Velocity Quick question When was the last time you actually visited a website to research something? If you're like me, AI pretty much does that work for you now? That, of course raises a new question for brands. If AI is doing the discovering, researching and deciding who or what is your website really for that shift in user behavior, the rise of AI bots becoming your most important new visitors is what my sponsor Scrunch is taking head on. Scrunch is the AI customer experience platform that helps marketing teams understand how AI agents experience their site. Where they show up in AI answers, where they don't, and what's preventing them from being retrieved, trusted or recommended. And it's not just visibility. Scrunch shows you the content gaps, citation gaps, and technical blockers that matter and helps you fix them so your brand is found and chosen in AI Answers. Now for our listeners, Scrunch is providing a free website audit that uncovers how AI sees your site, where there's gaps, and how you're showing up in AI versus the competition. Run your site through it@scrunch.com aidaily you've tried in IDE copilots. They're fast, but they only see local silos of your code. Leverage these tools across a large enterprise code base and they quickly become less effective. The fundamental constraint Context Blitzy solves this with infinite code context. Understanding your code base down to the line level dependency across millions of lines of code. While copilots help developers write code faster, Blitzy orchestrates thousands of agents that reason across your full code base. Allow Blitzee to do the heavy lifting. Delivering over 80% of every sprint autonomously with rigorously validated code, Blitzee provides a granular list of the remaining work for humans to complete with their copilots. Tackle feature additions large scale refactors legacy modernization greenfield initiatives all 5x faster. See the Blitzi difference at blitzi.com that's B L I T Z Y.com One more quick thing before we get back to the show. If you are a business leader who is thinking about how all of this crazy open claw and agent stuff can impact your business, I've got something for you. If you go to EnterpriseClaw AI. You can sign up to get more information about a new executive Sprint that we're going to be doing. That will help leaders inside companies figure out what the real challenges and opportunities of agents and agent systems like OpenClaw are going to be for your particular companies. That program will involve you learning, at least on a personal level, how to build agents and agent teams so that you have that basis of experience to then walk through a set of blueprints for the types of challenges you're going to face around things like security, governance and more. The first cohort is kicking off in March, so head on over to EnterpriseClaw AI to sign up for more information. Welcome back to the AI Daily Brief. The single big overarching theme that we've been tracking throughout this year is the sense among many the AI builders and the most enfranchised AI users that a key inflection point has been hit, that the transformation in the capability set of the latest generation of models, plus the tooling and products that we put around them, have in core ways, shifted just how much we can do with AI. Alongside that discourse has been an invigorated conversation about AI's potential impact on workers, particularly white collar workers. There are some, like Andrew Yang, who have long beat this drum, who are now getting even louder. A week ago, Yang wrote about the market's reaction to anthropic releasing plugins for functions like legal, finance, sales and marketing, and this week he paired it with a new piece called the End of the Office. He wrote, someone in my family had AI program a website for them. This week it completed in minutes what used to take a designer or a firm days of work. How many roles essentially consist of processing information and then presenting it to someone to make a decision. Now not only the process and report will be automated, but perhaps the decision as well. This will result in the great disemboweling of white collar jobs. Now it's not as though the job displacement conversation has ever been far from view when it comes to AI. And throughout 2025 it did accelerate at least a little bit. The loudness and vigor of the conversation tended to be tied to our sense of how fast agentic capabilities were evolving. Which is why, of course, this year, as we turn the corner on agents, the tone has gotten so much more dire. And yet the tone is important. We are still largely having this conversation on the basis of anecdote. It's Andrew Yang talking about his family building something that they would have had to hire someone for before and extrapolating that out to everything. The more urgent the conversation gets, the more important it becomes to move out of this paradigm of anecdotes and suppositions and theories and into the realm of actual evidence. One of the early attempts to quantify AI labor disruption came from the Yale Budget Lab and the Stanford Digital Economy Lab last August. The study was called Canaries in the Coal Mine and sought to map changes in hiring practices across different sectors to their exposure to AI disruption. Now, the study had a lot of problems which I covered in detail at the time. Chief among them was that the change in hiring practices began around the release of ChatGPT in November 2022, long before large companies were actually using the technology at any sort of scale. Still, the paper did show that whatever the cause, white collar hiring incontrovertibly had slowed down since then, particularly for younger workers. Now, one of that paper's authors believes he's seeing the first real signs of AI productivity showing up in the macro data. Now, at first blush, this might not be particularly surprising to you. Very few people are arguing that AI isn't extremely transformative in how we do work, or that it's not going to increase productivity. The conversation tends to be more about whether it does so so fast that it actually causes massive disruption. And yet, at basically every technology paradigm shift in the past, there has been a massive lag between when the technology comes on the scene and when the productivity increase actually shows up in the numbers. You might have heard this famous quote from Robert Sallow. You can see the computer age everywhere but in the productivity statistics. This was from 1987, and it came after a couple of decades of businesses investing in IT infrastructure. Now, at the end of the 90s, we did see a big productivity jump, but we haven't seen something in the last 20 years. In the wake of the Web2 mobile and social era, in fact, productivity growth has been lower than the historical pattern. If then AI is already showing up in the productivity numbers, that would in fact be a major break from the past and a major indicator of just how different this technology shift might be. In an opinion piece in the Financial Times, Stanford economist Erik Brynjolfsson argues that a recent revision to labor statistics implies the US is in the midst of a massive AI productivity boom. He points to the Bureau of Labor Statistics revising last year's job numbers down, which removed some 400,000 jobs from the statistics. Instead of 584,000 jobs created in 2025, net job gains are now believed to be just 181,000. Brynjelsen notes, however, that despite that downward revision, GDP figures for last year remained incredibly strong. We don't yet have the final numbers for Q4, but provisional statistics have growth at 3.7%. The Atlanta Fed's GDP now forecast is even higher at 5.4%, coupled with very strong growth at 4.4% in Q3 last year looks to be an incredibly strong year overall for the U.S. economy. Now, of course, we are talking in very macro terms, not at all getting into the lived experience of the economy for people, which is a whole different story. But this has some pretty meaningful implications for productivity stats. On this macro level, productivity is simply GDP divided by the number of workers in the economy. Thus when we got that downward revision of the workers, it implies much stronger productivity. Brynjolfsson estimates that last year's productivity growth will come in at 2.7%, which is almost double the average pace over the past decade, according to Brynjolfsson. This shift in the data would solve the conundrum put by Apollo chief economist Torsten Slok, who wrote in a weekend note referencing that same quote that I did AI is everywhere except in the incoming macroeconomic data. Maybe there is a J curve effect for AI where it takes time for AI to show up in the macro data. Maybe not. This J curve idea actually featured prominently in Brynjolfsson's research over the past year, and he believes that we may now be reaching an inflection point. In a paper from all the way back In October of 2018, Brynjolfsson and his colleagues wrote about the productivity J curve. Hilariously, the abstract of the paper focuses on general purpose technologies, or as they put the acronym GPTs. Yes, indeed, this was when the vast majority of people had never heard of OpenAI, much less interacted with one of their GPT models. In any case, Brynjolfsson and his colleagues write in the paper, general purpose technologies such as AI enable and require significant complementary investments, including co invention of new processes, products and business models, and human capital. These complementary investments are often intangible and poorly measured in the national accounts, even when they create valuable assets for the firm. We develop a model that shows how this leads to an underestimation of productivity growth in the early years of a new general purpose technology and how later, when the benefits of intangible investments are harvested, productivity growth will be overestimated. In his piece this week for the Financial Times, he hearkens back to this saying general purpose technologies, from the steam engine to the computer, do not deliver immediate gains. During this phase, measured productivity is suppressed as resources are diverted to investments. The updated 2025 US data suggests we are now transitioning out of this investment phase into a harvest phase, where those earlier efforts begin to manifest as measurable output. Brynn pointed to the slowdown in hiring for AI exposed industries from his prior work as further proof of this phenomenon. He concluded, we are transitioning from an era of AI experimentation to one of structural utility. We must now focus on understanding its precise mechanics. The productivity revival is not just an indicator of the power of AI, it is a wake up call to focus on the coming economic transformation One of the things that I expect that you'll see a lot, especially as people get louder about their assertions about how powerful AI has become, is that the chorus of people who want to say it's actually not all that good and everyone who says different is just a hype boy will get even louder. Sometimes those voices will come from people who get paid for making speeches and being the one token AI skeptic on the panel. But sometimes there'll be folks like the Apollo chief economist Thorsten Slok. What Erik Brindofsson then is arguing is that if this productivity enhancement is really showing up in the data, it undercuts a big part of those skeptical arguments. And it's not just Eric. Many other economic commentators share this read on the data as soon as the jobs revision hit economic opinion writer Noah Smith commented, this means there's actually a huge productivity boom underway. By the way, it's AI. Alex Imas, a professor at the University of Chicago, referenced an article he wrote in late January which said bottleneck tasks will slow down the emergence of AI gains in the aggregate data, but organizational restructuring, training, and improvement in tools will reveal the productivity impact sooner rather than later. After seeing these latest numbers, he commented, I guess sooner came pretty quickly. Not everyone, though, agrees. Economist Guy Berger posted, I'd be careful about drawing this inference from that data point may turn out to be true, but it's very thin evidence. Guy pointed to the actual revised statistics to see where jobs were being added and removed. By and large, this revision was about removing government workers from the Doge cuts, as well as counting additional layoffs in mining and logging, transportation and the manufacturing sector. Brynjelsen had used his own paper as evidence that the layoffs were happening in AI exposed industries, but at least for these numbers, that just doesn't seem to be the case now. However much this particular set of evidence actually points to AI disruption, the discussion is one that's clearly on the rise and the sense that US White collar sectors are having a hard time is doing nothing but growing, the Kobaisi letter wrote on Monday. US White collar hiring is extremely weak. There are now just 1.6 job openings per 100 employees in the professional and business services sector, the lowest in the last 11 years. This ratio has more than halved since 2021 and is even lower than the 2020 pandemic. Bottom Furthermore, the ratio of unemployed to job openings in the industry is down to 4%, nearly matching 2020 lows. This comes as total job openings in the sector are down 1.4 million since the March 2022 peak to 1 million, the lowest since May 2020. Over the same period, the hiring rate is down 1.8 percentage points to 4.2%, in line with levels seen during the 2008 financial crisis. Summing up they say the US white collar recession is accelerating Politicians on both side of the partisan divide are starting to speak up, with some even reacting to Matt Schumer's viral article Something Big Is Happening, Republican J. Abernulti said, I think it's something that's healthy to talk about. The Post says fundamentally we should be afraid because AI is going to be disruptive and there's going to be a lot of job displacement. That is something we know to be true. Obernolte has a master's degree in AI and has worked in the tech industry for three decades, so he's naturally skeptical that massive disruption will cause irreparable harm. He continued, the other thing that the Post says is people are going to have fewer jobs as a result of artificial intelligence. The historical record says that is absolutely not true. What isn't up for debate, though, is the idea that a major labor force disruption is coming one way or another, with Obernulti adding, there will be job displacement. We need to reskill the workers that are in industries with that job displacement and equip them with the skills that they need to succeed in other industries. We are going to need a social safety net because there will be people that fall through the cracks coming from a very different political base, senator Elizabeth Warren said, I'm deeply concerned about AI and what it's going to mean when people go out one day for lunch and come back and their jobs aren't there anymore and that that happens to millions and millions of people. Now is the moment when we need to be preparing ultimately to understand what's actually happening. It's just going to take a lot more data and understanding. In addition to that Financial Times piece. Brynjolfsson and his colleagues also recently circled back to their Canaries in the Coal Mine paper and wrote a follow up. Responding to some of the responses, Brynjolfsson tweeted Since our paper, there has been an explosion of research seeking to assess the relationship between AI and employment, specifically for early career workers in highly exposed occupations, where we found some of the most striking effects. While several studies have confirmed our initial findings, others have raised questions in two areas in particular. 1. Can the sharp increase in interest rates in 2022 explain the employment effects we observe better than AI exposure? 2. Is the overall timing of the employment effects consistent with AI exposure? They said in a follow up. Note that in brief, they find that one. While interest rates affect overall employment, existing evidence does not suggest they are a good explanation for the disproportionate decline in entry level hiring in AI exposed occupations. 2. They said, we do find suggestive evidence that when you include the broadest set of controls, the timing of the employment decline in AI exposed occupations becomes significant only in 2024. Some of the earlier declines are likely due to a combination of factors, not just AI. No one right now can claim to fully understand everything that's happening, but I'm encouraged that as we head early into 2026, we're seeing more work. Like the Haas AI study that found that people using AI actually spent more time on task, and that Brookings research that we talked about a few weeks ago that looked not just at which jobs were most likely to be disrupted, but which workers had the best capacity to adapt to displacement. At the end of the day, whether you are pessimistic or optimistic, and obviously you know which side of that I fall on, change is not only coming, it is here. Navigating it well is going to take everything we've got and for now that is going to do it. For today's AI Daily Brief, Appreciate you listening or watching as always and until next time, peace.