AI Agents Are Moving Into the Real World

0:00On the one hand, people continue to debate about whether there is actually a market for personal agents. Some think that this is one where Silicon Valley entrepreneurs just have it wrong, that people don't actually want optimization and productivity in their personal lives. Others argue that that's a reductive view and that for all of the tasks that people actually enjoy, like certain types of shopping, there are a million other small annoying tasks that they would love to have automated away. Increasingly, that debate is getting new evidence. And as agents like Muse and GrokBot pick up steam, they are moving off of our screens and into the real world.

0:35The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI.

0:44All right, friends, quick announcements before we dive in. First of all, thank you to today's sponsors, KPMG, Blitzy, Robots and Pencils, and Hyperagent. To get an ad-free version of the show, go to patreon.com slash AI Daily Brief, or you can subscribe and Apple Podcasts. And if you want to learn more about sponsoring the show, send us a note at sponsors at aidailybrief.ai. Now, one other quick note, there are a lot of new folks around these parts, and I'm trying to get a better handle on who you all are, what you're here for, and what you'd like to hear more and less of. If you go to the AIDailyBrief.ai website, there is a big link to a fall listener survey.

1:17It'll only take you a couple of minutes and would be incredibly valuable for me figuring out how to evolve this and potential other shows. So again, I would be endlessly grateful if you would take two minutes to do that fall listener survey. You can find it on our webpage, AIDailyBrief.ai. Anthropic has announced what they are calling a big AI breakthrough in the biology field. Less than a week after confirming the existence of an AI-powered wet lab in the Bay Area, Anthropic has said, Claw discovered a novel enzyme system with properties reminiscent of CRISPR, with only high-level direction from our scientists.

1:49CRISPR is gene editing technology based on an enzyme sequence discovered in bacteria. It's used to mechanically cut DNA and splice in new genes. Anthropics Discovery looks similar to CRISPR in the way it features an array of repeated genomes. They acknowledge that they don't yet know its function or level of significance, but these enzyme systems are relatively rare and have previously unlocked capabilities like cutting, copying, and inserting genes into DNA. They claim the discovery took 1 ,000 clod agents, 21 hours, and around 210 million tokens, so somewhere on the order of$10 ,000 of compute.

2:21CEO Dario Amadei, who has a PhD in biophysics, promoted the result on X. He wrote, It's easy to dismiss this as a one-off or curiosity, but we've repeatedly seen a pattern where AI performance in new intellectual domains goes from weak to superhuman in a matter of a few years. In 2023, models struggled to do math at the level of an average high school student. In 2024, they started to do well on math competitions for the best high schoolers in the country. In 2025, they started to solve minor open problems. In early 2026, more significant open problems. And in late 2026, they are beginning to solve the top few open problems in all of mathematics.

2:55We believe AI for biology is on a similar exponential trend. Amadei said the need to confirm significant results using physical experiments is what led Anthropic to set up their wet lab. He disclosed that Claude isn't quite at the point of autonomously controlling lab equipment, but that's the end goal. He also mentioned that it's only a safety level 1 slash 2 biolab, meaning it doesn't handle materials that are dangerous to humans. Continuing, he wrote, In Machines of Loving Grace, I wrote about AI's potential to cure most diseases in 5-10 years. A goal that sounds impossible, but one I believe is just barely possible if AI is applied to every stage of the pipeline.

3:28The first step is showing that AI can first help with and then drive biological discoveries. Claude's discovery of related prior work that goes back decades, beginning with systems like CRISPR and continuing with discoveries like the bridge recombinates and Viper in the past few years. I believe that we're at the very beginning of finding such systems and developing them into powerful tools for biotechnology. So what did people make of this? For some, it was way overhyped. Ravid Schwartz Ziv writes, The sad thing is Dario knows better. He was a PhD student. He knows the significance level of this result.

3:59He knows that if he had walked into Bill's office, his advisor, with, We found an interesting system, but we still don't know what it does, and said he was ready to graduate, Bill would have kicked him out of the room. But somehow, when the IPO is around the corner, this becomes AI is starting to drive biological discovery. Lucas Harrington is a bit less dismissive, but does add context. He writes, As someone who did this kind of genome mining work during my PhD, finding a weird cluster of genes and repeats is often the easy part. The hard part, and where the real discoveries come from, is figuring out what the system actually does.

4:30Anthropic does not yet know what this does. They've shown that the repeated array produces RNAs, but not what those RNAs do, what the RT does with them, or whether the system has any of the programmable properties that make the CRISPR comparison justified. He concludes,

5:03Pratham says, Don't misunderstand it as Claude curing cancer. But why it matters, finding tools like CRISPR used to take human experts months of manual searching. Here, AI did that work and found something real at a speed no human team could match. Still crazy. Basically, as Tanisha Abraham sums up, it's very cool and people should be excited, but it's also very, very, very preliminary. Next, we head over to a couple stories from the United Nations, where in a speech whose major purpose was rejecting globalist control of AI, President Trump has rebranded artificial intelligence to superintelligence.

5:35Speaking before the United Nations on Tuesday, Trump said,

6:01In other words, Trump added,

6:07It sounds much better. It is much better, and it's more accurate. Let's see if I have any power. Maybe I do, and maybe I don't. We're going to find out soon. Now, the memes were infinite. Replit CEO Amjad Massad wrote, Your intelligence? It's fake. It's artificial. Mine? It's super. Super intelligence. A.K.A. SI. Jvi Mausowicz, when he got done beating his head against the wall, lamented the concern of every AI safety advocate when he wrote,

6:41Meanwhile, getting back to the substantive part of the speech, Trump's objection to a quote, globalist regulatory scheme did not come out of the blue. Heading into the UN meeting, a group of 20 countries signed a declaration that welcomed initiatives to pace the frontier and called for mandatory independent safety testing, common standards, and the creation of an international AI institution by member states, which is able to, quote, set standards, enable verification, and convene states when capability thresholds are crossed. The call was led by Finland and supported by nations including France, Germany, and Singapore.

7:14UN Secretary General Antonio Guterres supported the call and encouraged, quote, external experts, including from frontier AI labs and AI safety institutes, to engage with the panel to help inform dialogue on such matters. The glaring hole in the proposal is that neither China nor the US were in support of a new international body to govern their development of the technology. During his speech, Trump said, We will only encourage superintelligence, not rein it in. We are going to watch it closely through the Department of Justice. The United States leads the world in superintelligence and will continue to do so safely and responsibly.

7:44White House science advisor Michael Kratios reinforced the point on Wednesday, telling the UN Security Council, Our primary responsibility is to ensure that continued advances in superintelligence benefit our citizens and safeguard their liberty and way of life. The frontier of intelligence is advancing rapidly. That is not a reason to pause its further development or to constrain it with new global governance structures. After staking America's claim to a long line of permissionless innovation that drove global growth, Kratios concluded, A prosperous future will not be secured by a global regulator.

8:13It will be secured by sovereign nations that adopt superintelligence, responsible companies that build it, and free people who refuse to be ruled by fear. Also on Wednesday, Sam Altman and Dario Amadei addressed the UN Security Council to explain how they think the council should think about AI risk. Sam Altman appeared live in New York and said, This moment is complicated as we feel both tremendous potential and very understandable anxiety at the same time. These models can help people discover new knowledge, new medicine, better healthcare, stronger education, and build more productive economies.

8:40They can be new tools for people to pursue dreams and ambition that today feel out of reach. Altman's entire speech was all about the dichotomy of AI, both the clearest danger and the biggest opportunity of the coming decades. He formed his suggestions around two major risks. First was the risk of RSI allowing frontier labs to build models they can no longer understand or control. Altman said, The industry must not accept too much technological risk just because the benefits are too great and they feel too important to slow down. He added, We need to understand what these systems are doing and have strong evidence that they will do what people intended, especially as they get very smart.

9:12It doesn't matter whether people put their risk of catastrophe at 10 % or 1%, none of these levels are remotely acceptable. Altman's second concern was concentration of power, commenting, A company or country that believes only it can be trusted with this technology can use that belief to justify almost anything else. Charting a middle path, Altman said, We don't want to fall into the trap of blind optimism. We also don't want to fall into the trap of doomerism either. While Altman didn't offer an opinion on global AI governance, Dario Amadei was far less coy. Calling in via video link, he said, I believe this is the most important global security issue facing the world today.

9:46No leader, no company, and no government can manage this alone. Amadei spelled out a step-by-step playbook for global AI governance. His prescriptions, as we have recently heard from him, include signing simple treaties with broad consensus like a ban on AI to develop biological weapons. Amadei then proposed the development of global evaluation and verification systems to allow states to verify commitments made by the labs in rival countries. Third, Amadei wants to see global standards for testing AI models for loss of control and misuse risks, as well as a global notification system. Emphasizing the gravity of the situation from his vantage point, Amadei warned the Security council that, quote, if managed poorly, I even believe that AI could be a risk to humanity as a whole.

10:24Anthropic on one pole of the safety discourse, the White House on the other, and open AI somewhere in the middle, trying to find the middle path. Now, if you are interested in that middle path, that's going to be something we talk about in this week's Long Read Sunday episode. However, for now, that is going to do it for the headlines. Next up, the main episode.

10:44Hello, everyone. One big change around AI is we've shifted our thinking from how we rank our pages to how do we become the source that AI trusts enough to answer with. At KPMG, they're seeing this firsthand. AI-generated results now surface answers directly, often without a single click. That's why they are increasingly focused on generative engine optimization, or GEO, structuring content so AI systems can retrieve it, understand it, and cite it as trusted authority. This is not just an SEO evolution, but a visibility mandate. And indeed, the GEO mandate from KPMG is simple. If AI is shaping decisions, your expertise needs to show up inside the answer.

11:21Read all about it at kpmg.com slash us slash GEO. Again, that is kpmg.com slash us slash GEO. Blitzy's deep code-based understanding unlocks the thing every roadmap owner cares about, shipping new features. Here's the truth about building inside a massive enterprise codebase. Writing code was never the bottleneck. Context is. Which system does this touch? Which contracts can't break? Which standards apply? Blitzy already knows because it reverse-engineered your entire codebase into a dynamic knowledge graph before feature work began. With that complete picture, Blitzy builds features end-to-end.

11:55Architecture, APIs, UI, and tests all validated against your existing systems. One Blitzy customer built an AI-native application from scratch with 100 % autonomous completion, Saving over 2 ,700 engineering hours. Features that respect your code base instead of fighting it. Stop letting your backlog grow faster than your team. Accelerate your roadmap at Blitzy.com. That's B-L-I-T-Z-Y dot com. The best teams don't have a single star carrying everyone else. They know their own strengths and each other's weaknesses and play to both. That's the team Robots and Pencils has built on purpose. Nobody there is grinding through busy work to pad a headcount number.

12:29People come for the hard problems and they stay because everyone around them is leveling up at the same time. In a market full of companies that are just trying to hire fast, that's worth a look. Check out robotsandpencils.com slash careers. This episode of the AI Daily Brief is brought to you by HyperAgent, where you run fleets of agents your team can manage together. Forget local agents and chat workflows waiting on your laptop to be prompted. HyperAgent deploys always-on agents in the cloud, doing real work across the tools your team already uses. Marketing agents turn competitor moves into landing pages.

13:00Sales agents enrich leads, draft emails, and updates the CRM. Ops Agent chases the paperwork and tracks the budget. Every agent has access to shared context and follows your rules about scope and approvals. It's time you add agents that feel like teammates. Hire yours at HyperAgent. Get$100 in credits at hyperagent.com slash AI Daily Brief.

13:22Welcome back to the AI Daily Brief. Today we've got everything from building websites while driving a car, to cute little mascots wearing smart glasses, because the theme of the show is AI moving into the real world. Now, if you've been listening over the past couple of weeks, you'll know that the arrival of GrokBot and more recently Muse, alongside a slew of other buzzy personal agent products like Instinct, have significantly increased the conversation around the place and resonance of personal consumer agents. And yet, despite all of these new products, there are some folks who are bringing a very deep baseline skepticism to this.

13:55Strategari's Ben Thompson has been talking about this across numerous interviews over the last month, basically arguing, as he put it to Patrick O'Shaughnessy, that consumers do not care about being productive. Now, in a recent appearance on TBPN, he went viral reiterating that exact point. He said, Tech companies are like, let us book your flight for you. I've had a personal assistant for 10 years. He does not come within an inch of booking my flights for me. That's crazy talk. I understand which ones are going to be delayed. I know which seat I want to be on. So the fact that this is continually a demo is hilarious to me.

14:26Number two, people love shopping. They don't want to be productive. They want to be entertained. One of the biggest forms of entertainment is shopping. This is part of the whole fast fashion thing. It's fun. Building up a head of steam, he continued, just because someone's complaining doesn't mean they're unhappy. People groan, oh, I have to research this. I have to go and do this. I'm so busy. And then you have an eager tech company executive say, oh, I have an assistant that can do it for you. And it's like, no, not only do I want to do it, I want to complain about it. I'm enjoying myself complaining right now, okay?

14:53Would you just let me be and live my life? Now, Ben is not alone in this sentiment. Earlier this week, Stay Sassy had a similar take, get a lot of traction. They wrote, The problem with all these AI personal assistant things is that people don't actually want to do stuff. You get that, right? Just like AI tools in the workforce, AI assistants don't actually abstract decisions for you. They let you accelerate decision making. So instead of making one decision a day and then lamenting you can't do more, AI assistants are going to let you make 10 decisions a day and act on all of them. A miracle?

15:22Nah, dog, a nightmare. Nobody except type A strivers who need Twitter fodder actually want that. After retelling a story about finding inspiration at college and then going home for the summer to see friends who didn't leave town, they commented, What I learned after many years of trying is that absolutely nobody wants to realize their potential. People want to hang out with friends and complain. That's how the world has worked for 10 ,000 years. This weather sucks. Can you believe the cows walked off? People don't want enablement. People don't want you to take all of their excuses. There's even a quote from legendary author Kurt Vonnegut from an interview back in 2005 touching on exactly the same topic.

15:55On leaving the house to buy an envelope, his wife suggested that he could just buy 100 envelopes online and never need to complete the errand again. Vonnegut continued, And so I pretend not to hear, and go out to get an envelope because I'm going to have a hell of a good time in the process of buying one envelope. I meet a lot of people, and see some great-looking babies, and a fire engine goes by and I give them the thumbs up. And, and I ask a woman what kind of dog that is. Of course the computers will do us out of that. And what the computer people don't realize, or they don't care, is we're dancing animals.

16:25You know, we love to move around. And we're not supposed to dance at all anymore. But what's the counterpoint? But just how correct is this nostalgic, romantic idea? What room and demand might there be for people to actually use personal assistants that are designed on the sort of productivity that Ben Thompson is dismissing? Investor Amy Wu Martin points out that trying to paint everyone with a single brush is inevitably reductive. Reposting Ben's soliloquy, she writes, Respectfully disagree. There are window shoppers and there are affluent consumers who hire a personal stylist or use Stitch Fix to save time shopping.

16:58There are travel point maxers and there are travelers who book via a travel agent because, again, time. Personal agents will not be for most people at the beginning. In fact, I'm guessing Muse retention will look terrible before its smile curves. But life optimizers will use them, assuming agents get privacy, fraud, etc. right. Fintech content creator Simon Taylor, meanwhile, agrees with part of Ben's premise, but thinks it's incomplete. He writes, Ben says, one, people don't want to be more productive, they want to be entertained, and two, shopping is entertainment. His point is solid. If you look at doom scrolling, Netflix watching, and shopping, you see a perfect consumer.

17:31The companies that fed the funnel won. Productivity apps happen but never take over the world. Those points are correct. But I think Ben is missing something big here. People are lazy. They want to delegate schlep work. They don't want to book the cycle class for my daughter by checking my calendar, finding the form, filling it in, confirming it with other calendars so they know too, then pay the$1 deposit for no-show prevention. And I might know where I want to book dinner, but picking a time that works for others, using the form, putting it in people's diaries. That's all schlep work I want to delegate.

17:58And users of Muse and Instinct will tell you it's magic. That's not shopping, it's buying things. It's the opposite of entertainment. Investor and entrepreneur Jill Gunter, meanwhile, says that this whole argument misses big chunks of the population. She wrote, I see this take all the time and I will continue to insist that if you don't think a lot of people want to be more productive or outsource admin, then you don't understand that there are 2 billion mothers in the world. Finally, Maduguru, who works on AI at Meta, puts it this way. Consumers don't have one relationship with doing things. As an example, they point out, browsing clothes can be entertainment for some, but hiring a roofer is miserable for most.

18:36Even with shopping, sometimes you want to browse for an hour, sometimes you want the right thing just delivered ASAP. The argument overall, the latent demand for agents that remove this friction for consumers is immense. And yet, unlike so much about the AI future, this is an argument that we actually can start to get some evidence for. And the newest piece of evidence is that right now, as I record, Muse continues to sit at the number one spot on the App Store. And even more than that, because there's all sorts of ways that a company with the size and distribution of a meta can drive an app to number one, even if it's not all that resonant.

19:08It's also getting really positive buzz from some unexpected areas. The New York Times' Eli Tan recently published a piece that was very surprising to many. It was called, I Gave My Life Over to Meta's AI Agent and Was Blown Away. Muse helped handle my dental insurance, booked dinner reservations, and created a podcast for me. It also required my most personal data and trust. Now, a big part of the article is about the challenge of the bargain that's inherent in making a personal assistant available, which is turning over so much highly sensitive and private data. But once the author did, they found, quote, Muse to be the most useful AI app I had ever used.

19:42So what did they actually do with it? Eli wrote, I had a few goals in mind. I wanted to see if Muse could bring me the personal empowerment that Mark Zuckerberg has often talked about. I also wanted to push Muse to see just how much AI agents can achieve and where they might fall short. Very quickly, I found that my agent could efficiently fill out online forms, a tedious part of using the internet. I could take a photo of a parking ticket and Ren would pay it for me, asking for approval before making the payment. My agent could also make phone calls using different personalities, and I found it to be helpful for dealing with customer service, such as with my dental insurance.