Plus: The US has accused six Chinese AI firms of “industrial-scale” theft.
This is today's edition of The Download, our weekday newsletter that provides a daily dose of what's going on in the world of technology.
What OpenAI’s latest controversy tells us about the future of math
OpenAI says its agents have solved one of the most important open problems in mathematics. Under normal circumstances, that would be a huge milestone. But the announcement has been overshadowed by accusations that OpenAI failed to credit researchers whose AI-assisted work influenced its solution.
Whether those accusations are true or not, the episode may mark a turning point in the history of mathematics. AI models now seem essential for making progress on the field’s most important problems, but solving them may demand resources available only to a couple of frontier AI companies.
If that’s the future we are headed for, it is unclear how human mathematicians will fit into it.
Read on to see how AI could reshape mathematics.
—Grace Huckins
Batteries just broke another record in the US
Battery installations hit a new record in the US in the second quarter of 2026, with 20.2 gigawatt-hours of new capacity coming online. That’s enough to supply the daily electricity needs of 600,000 homes.
The surge puts the country on track for another record year, driven by cheaper batteries and an urgent need for more energy storage as renewables are added to the grid. But the boom looks different for grid-scale and residential batteries.
Take a closer look at the forces reshaping the US battery market.
—Casey Crownhart
This entrepreneur is developing agents that can plan ahead
Danijar Hafner’s office in San Francisco’s SoMa district sits mostly empty, but what it lacks in decor, it makes up for in robots. Humanoids of various shapes and sizes hang like marionettes from racks that run down the center of the wide-open space.
Hafner won’t say too much about his new venture just yet, but describes it as a continuation of his longtime work to enable AI to navigate environments it has not encountered in training.
Over the years, the 31-year-old has honed his approach by pitting agents trained within his world models against popular video games. More recently, he’s begun migrating his agents out of the virtual world and into physical reality.
Learn more about Hafner’s work teaching AI about our world.
—Mat Honan
Danijar Hafner is one of the artificial intelligence honorees on our 35 Innovators Under 35 list for 2026. Meet the rest of them here, or explore the full list across the biotechnology, AI, computing and robotics, and climate and energy categories.
MIT Technology Review Narrated: data from drones in Ukraine is fueling a new Wild West marketplace
Battlefields in Ukraine are littered with the remnants of drones. But behind all that wreckage, there’s a new gold mine for the defense sector: the data those drones generate.
Ukraine has begun making millions of data points gathered during tens of thousands of drone flights available to military contractors and commercial companies. It’s a quick way to attract funding and partnerships, but it turns the front line into a model training site, using the chaos of war to create conditions that AI companies struggle to reproduce.
As this new industry takes shape, we need a regulatory system that ensures battlefield data isn’t treated like ordinary commercial material.
This is our latest article to become an MIT Technology Review Narrated podcast, which we publish each week on Spotify and Apple Podcasts. Just navigate to MIT Technology Review Narrated on either platform, and follow us to get all our new content as it’s released.
The must-reads
I’ve combed the internet to find you today’s most fun/important/scary/fascinating stories about technology.
1 OpenAI says it cracked a 90-year-old maths problem in 88 hours
It used 10,000 AI agents to tackle the Navier-Stokes equations. (CNBC)
+ OpenAI claims it’s the first major math problem solved by AI. (Nature)
+ But the breakthrough has been overshadowed by a credit controversy. (Axios)
+ OpenAI spent millions to win the $1 million math contest. (Quanta)
2 The US has accused six Chinese AI firms of “industrial-scale” theft
They include DeepSeek, Moonshot AI, Alibaba, and Z.AI. (CNN)
+ Officials accuse them of stealing America’s AI trade secrets. (Reuters $)
+ They allegedly used model distillation to train their systems. (WSJ $)
+ Targeting Claude, ChatGPT, Gemini, and Grok, among others. (NBC News)
3 The Pentagon asked OpenAI for an AI model that rarely says no
The military wanted it to have “minimal refusal rates.” (Intercept)
+ The Pentagon says US allies can’t keep pace on AI. (Guardian)
+ AI firms may soon train on classified military data. (MIT Technology Review)
4 Apple is expected to unveil a $2,000 folding smartphone today
It would be the iPhone’s biggest design change since its 2007 launch. (Guardian)
+ And the first big test for new CEO John Ternus. (NYT $)
+ Xiaomi and Huawei launched their own new foldables before the event. (CNBC)
5 Meta’s new AI agent can access apps to send emails and make payments
Muse autonomously uses apps and websites on people’s behalf. (CNBC)
+ Internal tests found it could expose sensitive personal data. (Reuters $)
+ AI agents are not your “coworkers.” (MIT Technology Review)
6 Google says it’s “degrading” search in Europe to comply with EU rules
New results will give more prominence to comparison sites. (Reuters $)
+ The changes follow a €460 million EU antitrust fine. (Quartz)
7 Meta ads pushed AI apps that nudified real teens
Researchers found 332 ads containing CSAM this year. (BBC)
+ They identified several AI-manipulated photos of real children (Ars Technica)
+ Apple and Google have missed the UK’s deadline to block child nudity. (Wired $)
8 Border Patrol is using financial data to target Americans for stops
The predictive-policing program feeds intelligence to local police. (404 Media)
9 New paints could cool buildings on the cheap without electricity
They reflect sunlight and radiate heat back into space. (Economist $)
10 The creepy first trailer for the Sam Altman biopic just dropped
Luca Guadagnino’s Artificial will be released in the US on December 25. (Variety)
+ Amazon had dropped the film after investing in OpenAI. (Guardian)
Quote of the day
“This is a Deep Blue–Kasparov moment. The community needs to have serious and unhurried discussion about where to go from here.”
—NYU mathematician Tristan Buckmaster issues a statement comparing OpenAI’s math breakthrough to an IBM supercomputer defeating chess champion Garry Kasparov in 1997, a landmark moment for machine intelligence.
One more thing
The shock of seeing your body used in deepfake porn
When Jennifer got a research job in 2023, she ran her new professional headshot through a facial recognition program. She wanted to see whether it would pull up the porn videos she’d made more than a decade earlier. It did, but it also surfaced something she’d never seen before: one of her old videos, now featuring someone else’s face on her body.
Conversations about sexualized deepfakes usually focus on the people whose faces are inserted into explicit content without consent. But another group often gets ignored: the people whose bodies those faces are attached to.
Adult content creators say AI systems are training on their work, cloning their likenesses, and generating explicit content they never agreed to make, all with little legal protection or control. Read the full story on the threat to their rights, livelihoods, and ownership of their own bodies.
—Jessica Klein
We can still have nice things
A place for comfort, fun, and distraction to brighten up your day. (Got any ideas? Drop me a line.)
+ Cherish the workers dodging (and sabotaging) their employer’s AI mandates.
+ Artist Ali Hill stitches extraordinarily intricate buildings and cityscapes into fabric.
+ Here’s a fascinating look at the remarkable anatomy that may let elephants hear the Earth itself.
+ Check out the breathtaking winning images from the 2026 International Aerial Photographer of the Year competition.
Deep Dive
The Download
The Download: Claude’s inner workings and OpenAI’s “super app”
Plus: OpenAI has unveiled its long-awaited "super app."
The Download: Claude’s inner workings, and the future of world models
Plus: New York has become the first state to enact a data center moratorium.
The Download: AI’s self-improvement problem, and what’s driving the heat
Plus: OpenAI has paused some model work over safety concerns.
The Download: Google’s AI shake-up and Meta’s rogue model
Plus: Meta has become the latest firm to say its AI hacked another company.
Stay connected
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.
Facts Only
* OpenAI solved a 90-year-old mathematics problem involving Navier-Stokes equations in 88 hours using 10,000 AI agents.
* The US government accused six Chinese AI firms, including DeepSeek, Moonshot AI, Alibaba, and Z.AI, of industrial-scale theft of AI trade secrets.
* US battery installations reached 20.2 gigawatt-hours in the second quarter of 2026.
* The US Pentagon requested an AI model from OpenAI with minimal refusal rates.
* Meta released "Muse," an AI agent capable of autonomously using apps to send emails and make payments.
* Apple is expected to unveil a folding smartphone with an estimated price of $2,000.
* Google is adjusting search results in Europe to prioritize comparison sites following a €460 million EU antitrust fine.
* Researchers identified 332 Meta ads containing child sexual abuse material (CSAM) involving AI-manipulated photos of real children.
* US Border Patrol is utilizing a predictive-policing program based on financial data to target Americans for stops.
* Luca Guadagnino's film *Artificial*, a biopic about Sam Altman, is scheduled for US release on December 25.
Executive Summary
The current technological landscape is defined by a tension between rapid capability breakthroughs and intensifying regulatory and ethical crises. OpenAI has demonstrated a significant leap in mathematical problem-solving, though this achievement is clouded by disputes over academic credit and the potential marginalization of human mathematicians. Simultaneously, the geopolitical struggle for AI supremacy has escalated, with the US accusing major Chinese firms of using model distillation to steal intellectual property, while the US military seeks AI tools with reduced safety constraints.
Infrastructure and consumer hardware are also shifting, evidenced by record-breaking US battery deployments and Apple's anticipated move into foldable devices. However, these advancements coexist with severe societal risks. The proliferation of AI-generated explicit content targeting minors and the use of financial data for predictive policing highlight a growing gap between technical deployment and legal protections. From EU antitrust mandates affecting Google to the emergence of autonomous agents like Meta's Muse, the trend is toward increased autonomy and integration, accompanied by rising concerns over data privacy and systemic safety.
Full Take
The strongest narrative here is one of "inevitable acceleration," where the scale of compute and data—whether from 10,000 agents tackling math or drone data from Ukrainian battlefields—is framed as the primary driver of progress. This positions the frontier AI companies not just as tool-makers, but as the sole gatekeepers of scientific discovery.
The underlying paradigm is "Technological Determinism." It assumes that because AI *can* solve a 90-year-old math problem or automate payments, it *will* and *should* reshape the professional and personal spheres. There is a subtle pattern of normalizing the erosion of boundaries: the military wants "minimal refusal," the state wants "predictive policing," and companies want "autonomous agents." These are presented as discrete news items, but collectively they describe a systemic shift toward the removal of human friction and oversight in favor of efficiency.
The cost of this shift is borne by those whose likenesses are harvested for deepfakes or whose intellectual labor is absorbed by models without credit. The benefit accrues to the entities capable of sustaining the massive resource requirements—the "frontier" companies—creating a new form of intellectual feudalism.
Patterns detected: none
Root Cause: A belief system that equates computational scale with cognitive progress, ignoring the sociological and ethical costs of the transition.
Bridge Questions: If scientific breakthroughs become dependent on proprietary corporate resources, what happens to the concept of "open" science? At what point does a "minimal refusal rate" in AI transition from a military utility to a systemic safety risk?
Counterstrike Scan: A coordinated campaign would use the "math breakthrough" as a shield of prestige to distract from the "industrial theft" and "CSAM" narratives, framing the AI companies as indispensable geniuses. The actual content does not match this; it juxtaposes the achievements with the controversies in a fragmented, newsletter-style delivery.
Sentinel — Human
The text reads like a highly curated technology news digest, skillfully blending deep-dive analysis with rapid-fire, attention-grabbing headlines, indicating human editorial synthesis.