← The AI Daily Brief: Artificial Intelligence News and Analysis
The AI Daily Brief: Artificial Intelligence News and Analysis
What the Pope Actually Said About AI
00:26:21
Loading summary
Transcript1 lines
- [00:01]
A
Today on the AI Daily Brief, what the Pope actually said about AI before that in the headlines An Update on the Mythos Rollout 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, KPMG Blitzy, Robots and Pencils and section. To get an ad free version of the show, go to patreon.com aidaily brief or you can subscribe on Apple Podcasts. And if you want to learn more about sponsoring the show, send us a Note@ SponsorsIDailyBrief AI at the end of last week, we got a quick update about Project glasswing. Now glasswing, you might remember, was the name that Anthropic had given to the initiative through which they rolled out their new Mythos model in a very limited capacity to around 50 partners. It was all about wrestling with the cybersecurity implications of the model, and Anthropic said the project has so far identified more than 10,000 software vulnerabilities of high or critical severity. Most partners, they wrote, have found hundreds of severe vulnerabilities in a little over a month using the model. Testing the model's accuracy on open source repos found less than a 10% false positive rate, wrote Anthropic. Progress on software security used to be limited by how quickly we could find new vulnerabilities. Now it's limited by how quickly we can verify, disclose and patch the large number of vulnerabilities found by AI. Anthropic noted that industry standards demand that bugs are not disclosed for three months after identification, meaning that the comments are vague. At the same time, the report highlighted several public statements. Mozilla recently commented that they had found and fixed 271 vulnerabilities, which was more than 10 times the amount they found using Opus 4.6. Palo Alto Network's most recent release featured five times as many patches as well. The report also noted that some in the finance sector are also using the model for real time fraud detection. For one firm, Mythos was able to detect and prevent a $1.5 million wire transfer to a threat actor who compromised a customer's account. Now, one of the big criticisms of the Mythos rollout has been its exclusivity with the model provided to only a few dozen top US Companies. And in addition to the podcasters and power users of X who have been frustrated that they can't play with the new model, the frustration apparently extends to global policymakers as well. Earlier this month, European officials called in Anthropic to make Mythos available to their financial sector, with similar calls having come from the uk, Canada and governments across Asia. The Australian government held high level talks with Anthropic executives last week in an attempt to gain access. Anthropic, for their part, says that they are currently working with critical partners, including the US and allied governments, to expand Project Glasswing and said that while they look forward to releasing a Mythos class model to the public, that timeline is unclear at present. They wrote, no company, including Anthropic, has developed safeguards strong enough to prevent such models from being misused and potentially causing severe harm. This is why we have yet to release Mythos class models to the public. Testing catalog did spot a model called Claude Mythos one preview being prepared for release in Claude Code and Claude Security. Their assumption is that this won't be the public release, but rather broader access for trusted partners to use the model in Anthropic's harness. So where this all stands first, Mythos has seemingly changed the nature and speed of cybersecurity work. While these partners are reporting a 10x increase in the number of vulnerabilities detected, the bottleneck now moves to actually designing and deploying bug fixes. In their update, they wrote, several maintainers have told us that they are currently severely capacity constrained and some have even asked us to slow down our rate of disclosures because they need more time to design patches. The bottleneck they continued in fixing bugs like these is the human capacity to triage, report and design and deploy patches for them. Box CEO Aaron Levy suggested then that the impact of Mythos is ultimately going to be a microcosm for the deployment of AI tools more generally, he posted. This is precisely an example of why engineers don't go away, ever. We've made it far easier to create and find security issues, which means the new bottleneck is in our ability to actually review, respond to and fix the issues. Far from AI magically solving all of this, there is still major triage work and human judgment required to do the follow on work to actually protect systems. As a result, we're about to enter a security engineer Boom. Jayvon's Paradox all over again. Now, speaking of Anthropic's partners in the government, the White House has approved a secret $9 billion budget request for intelligence agencies to build their own inference cluster. The New York Times reports that the CIA and NSA made the budget request because they're unable to run the latest AI models on their current classified systems. The funding will go towards the purchase of Nvidia Blackwell chips and supporting infrastructure, with the report noting government concerns about a shortage of Blackwell Compute and the need to run Mythos class models in classified settings. The New York Times suggested that the Pentagon and intelligence services have fallen behind in the AI buildout after failing to allocate enough funding in recent years. After putting the issue to the White House for comment, the New York Times received what they characterized as a sharply worded statement. A White House spokesperson said sensitive national security deliberations are conducted with the seriousness they demand, not leaked to reporters and repackaged through selectively sourced, unverified claims designed to drive headlines rather than truth. The fact is, the United States is leading the world in technology and is well prepared to deal with a variety of issues that may arise. The reporting also discussed progress in contracting between intelligence agencies and frontier AI companies when OpenAI signed their agreement with the Pentagon in March. That contract specifically excluded the nsa. Sources said that White House Chief of Staff Susie Wiles has approved the use of anthropic models at the nsa, despite the Pentagon designating Anthropic as a supply chain risk. They added that Anthropic and the government are finalizing a contract to allow the NSA to use anthropic models in classified settings, which could help end the dispute with the Pentagon. Intelligence officials said they hope the contract with Anthropic will also pave the way for an agreement with OpenAI that includes spy agencies. Interestingly, it seems the intelligence agencies are more flexible on guardrails than the Pentagon. The dispute with Anthropic kicked off over a limitation on using AI for autonomous weaponry and domestic surveillance. The Pentagon argued these limits could jeopardize active military operations, demanding that the contracts instead allow AI to be applied for any lawful use. The NSA and CIA, meanwhile, are technically prohibited from gathering intelligence within US borders, so are reportedly more comfortable excluding this language and agreeing to outright bans on using AI for domestic surveillance. Now our next story is on the face of it just about price and token access, but clearly also has a geopolitical bent as well. Deepseek has made their deep discount permanent as they close in on a major funding round. DeepSeek's V4 model launched in April with a 75% discount on tokens. That made the model 1/7 the cost of Opus 4.6 and a quarter the cost of GPT 5.4, which were the latest models from each of the major labs at the time. Since then, Anthropic and OpenAI have each raised prices on their updated models, with the discounted price now just the price. Deepseek is serving a model that's roughly equivalent to Opus4.5 for $0.44 per million input tokens and $0.87 per million output tokens. In addition, Deep Seq management have told investors that they're going ahead with a $10 billion funding round, which will be the first time Deep Seq has taken outside fund, ending with the round reportedly valuing the company at $45 billion. The Chinese government's AI industry investment Fund is expected to participate alongside multiple large Chinese companies including Tencent and JD.com founder Liang Wen Fang has told investors that the focus won't shift to monetizing Deep Seq's technology. He said that the firm will continue to develop open source models and push the boundaries of AI research with an end goal of achieving AGI. If the round closes at these numbers, it will be a new record for the Chinese venture ecosystem, bloomberg analysts wrote. Asia's AI models are decoupling from the US as they shift towards a token based economy. China is leveraging low power costs and a huge developer pool to treat AI tokens as tradable assets. The quote unquote industrial approach to AI is being fueled by a surge in one person firms using tokens to raise productivity. What will be interesting to see is as token constraints and token budgets get more pinched in the us, will companies start to look to these open source models as a viable alternative? Certainly, if you just read the tea leaves of how young startups are doing it, the answer is likely to be yes. Lastly today a little bit of model news to close us out. GROK might still have a few iterations left as Elon Musk announces a successful new training run. Now I think it's a fair assumption that at this point Elon is at least orienting towards a new role as SpaceX becoming an AI cloud company. But at least for the moment, it appears that Xai, which is now SpaceX AI, isn't fully done with training their own models. Elon announced that a foundation model called Grok V9 Medium has finished training. The model has 1.5 trillion parameters with Musk writing evals look good. A lot of cursor data was added in supplementary training and there is more to come. Fine tuning is underway and reinforcement learning begins in a few days. Two to three weeks to public release. This will be a major improvement over the 0.5 trillion V8 small that currently serves all GROK production traffic, especially for difficult coding tasks. In addition XAI has rolled out access to Grok Build for Super Grok and X Premium. Grok Build is xai's competitor for Codex and Claude code and was previously only available for Grok Heavy users, which is the premium $300 a month. T People seem pleasantly surprised. Justin Schroeder of dmux says, Grok Build is surprisingly good, much better than I expected. The harness is solid, models are very fast, and it already seems less buggy than Claude now. I'm still not exactly clear how this is all going to shake out in terms of strategy, but for the moment at least, it appears that we will continue to get new Grok updates. That, however, is going to do it for the headlines. Next up, the main episode all right folks, quick pause. Here's the uncomfortable truth. If your enterprise AI strategy is we bought some tools, you don't actually have a strategy. KPMG took the harder route and became their own client. 0 they embedded AI and agents across the enterprise how work gets done, how teams collaborate, how decisions move not as a tech initiative, but as a total operating model shift. And here's the real unlock that shift raised the ceiling on what people could do. Humans stayed firmly at the center, while AI reduced friction, surfaced insight, and accelerated momentum. The outcome was a more capable, more empowered workforce. If you want to understand what that actually looks like in the real world, go to www.kpmg.us AI. That's www.kpmg.usa 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 codebase and they quickly become less effective. The fundamental constraint context blitsly 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 Blitzi to do the heavy lifting, delivering over 80% of every sprint autonomously with rigorously validated code, Blitzi 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 thing I keep seeing in enterprise AI companies hedging across every cloud, every model, every framework, or paying a GSI for a pilot that never ends the team's actually shipping They've picked a lane and they move fast. That's one of the reasons I like today's sponsor Robots and Pencils. They've gone all in on aws. They're an advanced tier and AWS pattern partner, and they ship production AI coworkers in 45 days. That's led to them doing some of the more interesting work I've seen on AI coworkers. And by that I'm not talking about chatbots. I'm talking about actual agentic systems that sit inside a business architecture and do real work. That kind of focus matters if you're an enterprise leader trying to get something real into production, or an AWS rep trying to move a customer from interested to deployed. Request an AI briefing at robotsandpencils.com One conversation with robots and Pencils and you'll know Here's a harsh truth. Your company is probably spending thousands or millions of dollars on AI tools that are being massively underutilized. Half of companies have AI tools, but only 12% use them for business value. Most employees are still using AI to summarize Meeting notes if you're the one responsible for AI adoption at your company, you need section Section is a platform that helps you manage AI transformation across your entire organization. It coaches employees on real use cases, tracks who's using AI for business impact, and shows you exactly where AI is and isn't creating value. The result. You go from rolling out tools to driving measurable AI value. Your employees move from meeting summaries to solving actual business problems, and you can prove the ROI. Stop guessing if your AI investment is working. Check out section@sectionai.com that's s e c t I o n a I.com welcome back to the AI Daily Brief. Today we have something that is very different from our normal coverage. Obviously, at this point it's not surprising that AI is increasingly not just a technology issue or even an economy issue, but a larger political and social issue as well. And yet I've rarely seen the amount of social discourse around AI as a global societal force, as we've had in the last 24 hours in the wake of the release of Pope Leo XIV's Magnifica Humanitas. Magnifica Humanitas is what's called an encyclical, which is a major teaching letter from the Pope. They actually go all the way back to the early church, when important teachings were passed around Christian communities as actual physical letters. Encyclicals in general are not new doctrine. More often they interpret existing Catholic teaching in light of new circumstances and for popes in the modern era, they become an increasingly important tool as the Church responds to everything from industrialization to globalization. And now, with Leo's first encyclical to AI. One of the most famous of these is the Rerum Novarum from Pope leo XIII in 1891, which addressed labor capital, wages, workers rights, and industrial society at the height of the Industrial Revolution. They're not particularly common. Recent popes have tended to publish an encyclical every two or three years, and the first encyclical is often seen as important because it can signal some of the major themes of a pope's pontificate. Now, for those of you who are not familiar with Pope Leo, he was elected in May of 2025 and immediately identified of AI as the central social question of his papacy. He named himself as a tribute to Leo XIII, that same pope from the late 1800s who led the Vatican through the Industrial Revolution, Leo XIV, during his first address last May, said, quote, in our own day, the Church offers to everyone the treasury of her social teaching in response to another industrial Revolution and to developments in the field of artificial intelligence that pose new challenges for the defense of human dignity, justice, and labor. Now, my read broadly is that as much as people try to put the Pope in one bucket or another as either too antagonistic towards AI or not antagonistic enough, the central theme throughout all of his discourse on AI has been less about trying to designate it as good or bad, but rather acknowledging it as a pivot point in both history and in faith. Now, the other things that are worth noting about Pope Leo is that, A, he's the first American Pope, and B, he's also one of the first popes to actually use modern technology. He has a cell phone and a computer. He has an Apple watch, he watches cable tv. He's of the modern world in a way that the popes before him just simply were not. And so yesterday we got Leo's first encyclical, which again was called Magnifica Humanitas subtitle on Safeguarding the Human Person in the Time of Artificial Intelligence. The first thing that I will note, and the most important thing probably that I will say in all of this, is to echo AI policy Reacher Miles Brundage when he writes, my main encyclical take is that folks should just try to read it. Most of the takes are bad, as in most social media takes are bad, not most of the encyclical takes are bad. Obviously, regular listeners will know that I try to be as nuanced as possible and provide as wide an array of legitimate takes as possible. But even in this case, as you will see, I am going to put words in the Pope's mouth and ascribe what I believe is the central point of the whole piece, despite the fact that the thing is longer than a lot of books. Now, I will say that if the mark of a nuanced piece is lots and lots of people from lots of different sides of an argument being angry that it didn't do more to support their argument, then Pope Leo succeeded spectacularly, wholly deciding to not engage with anything in the actual text of the piece. Tim Nick Gebru instead focused on the fact that a representative from Anthropic was invited to give comments as part of the day's ceremonies, saying, in case you thought religious leaders were going to advocate for your rights and not the dudes about to make billions with their ipo, I guess it was only a matter of time when the religion of effective altruism worshiping so called artificial general intelligence merged with entrenched religious institutions on the other end of the spectrum. Lots of folks cherry picked a quote and extrapolated an entire position and set of assumptions from it. One of the quotes that came up most often for people on the other side of the argument was this. The pursuit of greater profits cannot justify choices that systematically sacrifice jobs because the human person is an end, not a means, and the economic order must remain subordinate to human dignity and the common good. Entrepreneur Blake Scholl wrote, bad take from the Pope. Tech revolutions tend to eliminate some jobs while creating others. If we cling on to jobs, we'd still be plowing fields by hand out of fear of disruption. The other thing that I saw a lot of with the encyclical is sort of the inverse of what we just heard with people being angry that it didn't support their position enough to instead cherry picking how it supported all of their positions. Connecticut Senator Chris Murphy wrote, really important AI threatens to undermine the basic building blocks of humanity as it seeks to replace our most basic functions like creativity, friendship and critical thinking. Chief water use lie propagator Karen Howe wrote, the Pope has weighed in. AI isn't magic. It is produced by a currently deeply exploitative extractive supply chain. It challenges human dignity. It could fuel a new colonialism, and among the greatest challenges of our time will be to redirect its path of development. One of the strangest but also kind of most revealing frustrations came from Dean Ball, who was, to be frankly, a little overly reductive about his position, upset that the Pope didn't grapple more with the possibility of AI cognition or, frankly, even personhood. He wrote a lot of tweets about it, but one of them read the reality of AI cognition is the central challenge the Church and all of us will have to grapple with over the coming decade. And this encyclical, with its axiomatic denial of AI cognition, is a punt of the highest order. Now, as some pointed out, Dean's critique kind of contains within it an axiomatic assumption of AI cognition that many, many people disagree fairly vociferously with. Although I think a lot of people's bafflement about this argument was better summed up by Jonathan Lytle, who wrote, Guy is extremely mad that the Pope of the Catholic Church believes in the soul, apparently. Or Josh Barrow, a little baffled by people who are upset that the Pope doesn't think a machine can be insold. What do you expect the Pope to think about that? And the reason that I'm hanging on this point is that I think it actually does get to perhaps the most central discourse in the piece. And so I want to give Dean the last word before we dig in a little bit deeper, where he responded to these critiques by saying, some think I want the Pope to insole AI or acknowledge AI feelings. I don't. What I want is for the Church to contemplate what humans should do, as we are eclipsed as the smartest entities on the planet, at least for many reasonable people's definitions of the word smart. Now, before we get into that central point, I want to take a big step back. Fascinatingly, I think the group that on average had the best read of the underlying tone and intent of the piece were Catholics themselves. Christopher Hale wrote, anyone who thinks Pope Leo XIV is an AI doomer either hasn't read the encyclical in full or or doesn't understand Catholic theology. The Pope is an AI realist. He knows its growth is inevitable. He just wants to ensure it's always in service of the human person. The Catholic SAT account on Twitter writes, the TLDR is AI is not inherently evil, but is never neutral and carries risk of power, concentration, inequality, and loss of human dignity. The Church offers principles for discernment rather than blanket rejection, and at a high level. That's what I got reading the entire piece as well. It's a document that very much positions us at the beginning of the beginning. To the extent it provides warnings, it's warnings for what could be. It's much more an exploration of what we should think about and discuss as opposed to specific things we should change. But to the extent that it does carve out positions, it's about the foundational principles that Leo believes we should bring into the discourse, rather than a strict delineation of the conclusions that those principles should lead us to. And perhaps unsurprisingly given that the subtitle is Safeguarding the Human Person in the Time of Artificial Intelligence, the core plea, in fact, of the piece is, in my estimation, to keep real living humans, not artificial intelligences nor anonymous market forces, as the barometer of our success with this and any technology. That line about the pursuit of greater profits, not justifying choices that systematically sacrifice jobs, is about the ever present tension between, on the one hand, markets as unfeeling, efficient allocators of scarce resources, and on the other hand, markets as ultimately servants of human wants and needs. AI is of course, reigniting this tension in a massive way. But that fundamental tension at the core of who markets are supposed to serve is and has been ever present. Now, when it comes to the Pope's argument about the potential for a new form of colonialism, it is not, I don't believe, the argument that Karen Howe paraphrased when she said that the Pope had weighed in and decided that AI is currently deeply exploitative. Instead, the piece recognizes that when a new resource is identified, the potential for exploitation and new power dynamics based on that resource become something that we have to pay attention to. The piece recognizes that people's data has become extremely valuable, ultimately a new type of power to mine, and that given that we are already seeing how much aggregation of that data can lead to economic power, we need to have a conversation, as the Pope puts it, about how to ensure that shared knowledge becomes a true common good rather than an instrument of dominance. Now, to me, reading that paragraph, which is 178, by the way, and more about the potential misuse of health data, the Pope writes, those who control the health data of entire peoples, often collected under the pretext of aid, research and innovation, possess a structural leverage over the future, for they can shape needs in markets. They can decide before others to whom medicines, investments and protections will be allocated. It's less an indictment of today and more a warning about tomorrow. And yet, more than anything else, to me, the central argument that underpins the entirety of Magnifica Humanitas is that human value cannot be reduced to an intelligence benchmark. That even if and when AI becomes quote unquote, smarter than us, it will remain categorically different from us. And in part, the reason for this is a rejection of the idea of the foibles and problems and challenges and limitations of humanity as inherently problems to be fixed. One of the most contentious paragraphs for people like Dean Ball, who have some belief in AI cognition is paragraph 99, where Leo writes, we must avoid the misconception of equating this type of intelligence with that of human beings. These systems merely imitate certain functions of human intelligence. In doing so, they often surpass human intelligence in speed and computational capacity, offering tangible benefits across many fields. Yet this power remains entirely tied to data processing. So called artificial intelligences do not undergo experiences, do not possess a body, do not feel joy or pain, do not mature through relationships, and do not know from within what love, work, friendship or responsibility mean. Nor do they have a moral conscience, since they do not judge good and evil, grasp the ultimate meaning of situations, or bear responsibility for consequences. They may imitate language, behavior and analytical skills, or even simulate empathy and understanding, but they do not understand what they produce, for they lack the effective relational and spiritual perspective through which human beings grow in wisdom. Even when these tools are described as capable of learning, their way of doing so is different from that of a human person. It is not the experience of those who allow themselves to be shaped by life and grow over time through choices, mistakes, forgiveness, and fidelity. Rather, it is a form of statistical adaptation based on data and feedback, which can be very effective but does not imply inner growth. As much as this may come as a shock to the researchers and technologists building AI, this is not going to be a controversial take among the vast majority of people, including people who use AI every single day. For most, the idea that even if AI can do math or write things faster and even better than us does not change the fact that it remains fundamentally different to us in ways that should be enshrined in our policy and positioning towards it. Now, what I will say is that as AI gets more advanced, there will be more debate to be had. Here, people who currently feel right now as I just argued, that AI being quote unquote smarter than humans doesn't make it of equivalent value from a societal and policy perspective to humans. My guess is that in the future more people will have more questions about that than they do now. In that way, I think the magnifica humanitas is a flag planting. It's a preparation for the period where that conversation heats up. Closely related, I think, is paragraph 118 where Leo writes, our relationship with life seems to be in crisis. Today, everything that appears as a limit incapacity, illness, old age, suffering, vulnerability tends to be seen primarily as a defect to be corrected, rather than as a reality through which our humanity matures and opens itself to relationship. And yet we must remember that humanity flourishes not despite limitations, but often through them. Now, from there the argument veers a little bit more into the theological than I want to go into in the context of this particular episode. But in many ways this pairs with the denial of the specialness of AI intelligence as central pillars in a fundamentally contra transhumanist type of argument that the encyclical is setting a foundation for. Now, obviously this debate is going to get nothing but louder, and in that I think Boyan Tungus might have it right when he writes the most important thing about Magnifica Humanitas is that it exists. Challenges posed by AI are real. They will only increase and they will have a massive impact on all aspects of human life. The fact that the Catholic Church takes this matter very, very seriously, to the point that Pope Leo decided to dedicate his first encyclical to it, is very laudable and encouraging. It shows urgency, centrality and importance given to this topic by one of the biggest and most visible global institutions. I hope that many other institutions, especially governments and professional organizations, start approaching matters of AI with the same focus and dedication and high level urgency. And I think that that's right. And I will only add as we conclude that despite some of the throwaway sound bitey type takes I shared earlier in the show, in general, the discourse and discussion around this particular piece has been so elevated compared to the average mud fest that is the discourse on AI and frankly any big political issue these days. Like I said at the beginning, I highly encourage you to take some time to read the whole thing or at least drop it in your AI of choice and have a conversation about it. For now, however, that is going to do it for today's AI Daily Brief. Appreciate you listening or watching as always. Until next time, peace.