↵ select ↓ ↑ navigate esc close

Daily Briefing: November 1, 2026, morning

Daily Briefing ·

Listen: https://blossom.buildtall.systems/f7af27204ef7f6919a928aeed5252c0b0c4fd3451b3c8078e733e6a63ab33676.mp3


The morning briefing for November 1, 2026, covers thirteen articles. The dominant currents are artificial intelligence adoption and reasoning, data-center politics, space exploration, and a major cryptocurrency theft response.

The first group concerns the spread of artificial intelligence through commercial products and development pipelines. Apple CEO says the company doesn’t ‘want to ship AI for AI’s sake’ is a short TechCrunch report. It states that Apple's new CEO defended the company's current position in the AI landscape, and provides no further detail. NVIDIA Updates Rubin GPU Specs as First Production Hardware Arrives is a brief ServeTheHome item. It says that as production hardware starts shipping, NVIDIA has updated some of the specifications for its Rubin GPU platforms.

Xbox Manager: AI Already Being Used In Xbox Game Development - And Throughout the Game Industry is a Slashdot report drawing on Engadget. It describes claims by Brandon Stirrup, a strategy manager at Xbox, that Xbox is already using AI and that developers who say they are not are merely pretending. Stirrup deactivated his social media account shortly after the posts, and the article notes that his statements were archived elsewhere. The piece situates the claim inside Microsoft's broader AI position and the recent formation of Xbox Labs, which was announced partly in response to a rising trend of AI-coded game ports but did not explicitly confirm AI use. It lists developers that have used AI or pledged to do so, including Capcom, Square Enix, Sega, Nexon and Activision, and describes Epic Games and its chief executive Tim Sweeney as particularly up-front and aggressive in promoting AI. At the same time, the article identifies substantial pockets of opposition, concentrated among smaller independent studios without shareholders to impress, including Finji, the publisher of Tunic, ConcernedApe, the developer of Stardew Valley, and Strange Scaffold. It says the division extends to players: some accept AI as part of games going forward, while others pledge to avoid games that use it because of environmental costs, labor issues or aesthetic preference. The report notes the popularity of tools that make Steam's AI disclosures more visible and grassroots labels certifying games as AI-free. It closes with the claim that many players and developers justify AI in code generation even when opposed elsewhere, and quotes Stirrup as saying that gaming is the only industry slow-walking the transition because users have negative sentiment, and that gaming is the only industry pretending to not use AI.

A second theme is the collision between data centers and local politics five days before the midterm elections. Judge allows NAACP lawsuit over xAI data center pollution to move forward reports from TechCrunch that an NAACP lawsuit accuses the SpaceX-owned AI company of violating the Clean Air Act. Data centers are now a midterm issue, even in Brooklyn is a companion TechCrunch item. It says that at a Brooklyn midterm party five days before Election Day, every conversation, from a council member's fireside chat to a debate about the Democratic Party, kept circling back to AI and data centers.

Two longer articles turn to space and the solar system. Computer Simulations Suggest Building Blocks of Life Didn't Arrive on Prehistoric Comets is a Slashdot article reporting on a CNN story. It examines a study published Wednesday in Scientific Reports that challenges a long-standing theory that comets were crucial in kick-starting life on Earth billions of years ago. The popular theory dates back to at least 1961, with earlier versions postulated in 1908. The new computer simulations suggest that fragile organic molecules supposedly carried by icy objects would have been destroyed on impact.

The study's lead author, Richard Anslow, conducted the research while completing his doctorate at the Institute of Astronomy at the University of Cambridge. The team used the latest models of comet formation and evolution to simulate the early planetary period one hundred thousand times. Early in its history, more than four billion years ago, Earth may have experienced a period of heavy bombardment from more than one hundred thousand comets larger than one hundred meters. According to Anslow, the typical velocities of these comets when they impacted Earth meant that very few, if any, would have been able to deliver intact organic molecules. Fifty-four percent of the simulations showed no impact at all, while the remaining forty-six percent had at least one collision, and only a tiny minority returned up to three collisions.

Anslow cautions that a successful impact cannot be ruled out confidently, but such a fluke would have had to occur in a perfect location that already met several important conditions, including the presence of liquid water and other chemicals needed to support life. Even then, the article notes, important chemicals such as hydrogen cyanide would have been unlikely to survive impact, and would have dissolved in an ocean impact. The study also points out that many other precursor molecules are needed, such as compounds containing phosphorus and sulfur, and these are not routinely found in comets.

The article reports that an alternative cosmic source of organic material could have been asteroids, which typically struck Earth more slowly and in greater numbers. Anslow points to recent findings on samples from the asteroids Bennu and Ryugu, which contain the majority of protein-building amino acids. The study also raises the possibility that ancient Mars might have been better suited to preserve hydrogen cyanide from comet impacts, because Mars is smaller than Earth and farther from the sun, so typical impact velocities are much lower. That idea adds another dimension to the long-standing question of whether the red planet once supported life. Queenie Chan, an associate professor of Earth sciences at Royal Holloway University of London, who was not involved with the study, adds that hydrogen cyanide may have undergone reactions before the icy bodies entered Earth's atmosphere, potentially creating biologically important compounds that are less volatile and more likely to survive.

Isaacman outlines Artemis III multi-launch checkout sequence and Blue Origin priority is a NASASpaceFlight.com article based on a wide-ranging interview with NASA Administrator Jared Isaacman. It describes the planned architecture and contingencies for the Artemis III multi-vehicle campaign. Isaacman emphasizes that the mission will require a sequenced series of launches involving three different vehicles and providers, and that NASA will only commit the Space Launch System if Blue Origin's lander is already in orbit.

The current plan begins with a New Glenn launch carrying Blue Origin's Blue Moon lander. That vehicle is expected to have the ability to loiter in orbit for roughly thirty days, allowing time for checkouts before the crewed vehicle arrives. The lander flight would be followed by the Space Launch System and Orion crew launch, and then by a Starship launch. Isaacman describes the three-launch, three-provider profile as unprecedented for a NASA crew mission and a reflection of the multi-launch campaigns expected for future large-scale space infrastructure. He states that NASA will not launch without Blue Origin in it, and that Blue Origin will launch first, perform checkouts, and then NASA will commit SLS.

The article reports that Artemis III is intended as a complex low-Earth-orbit checkout and rendezvous campaign, exercising the three major flight elements together, gathering operational data and building experience ahead of a later crewed lunar landing attempt. It is explicitly not a single-launch profile. Isaacman frames this as a deliberate departure from the Apollo model and from earlier Artemis concepts that assumed a more tightly coupled architecture. He says the approach reflects the realities of available hardware and NASA's desire to gain experience with multi-vehicle rendezvous and docking before a subsequent surface landing attempt. He also draws a direct line to the idea of autonomous assembly in orbit and fleets of fission-powered spacecraft, which would themselves require many launches to build.

On the comparison with Apollo, Isaacman says that Artemis III is not an Apollo 9, just as Artemis II was not an Apollo 8. Instead of attempting to turn the pad and hit a landing in the same time frame, NASA aims to build muscle memory, launch the big rocket, and extract as much as possible from the mission. Once SLS and Orion are committed and on orbit, the agency intends to obtain as much value as possible from the Blue Moon element. Starship is expected to contribute high-definition imagery, controllability testing, and rendezvous and docking practice during the roughly two-week period that includes two planned rendezvous events. Isaacman says the primary operational learning is expected from the Blue Origin side, because that team has not previously kept people alive outside the New Shepard suborbital environment for extended periods.

On schedule drivers, Isaacman identifies the lander as the primary remaining bottleneck. He says there is not yet a fully assembled Blue test article, while adding that Blue Origin is making substantial progress. At the same time, he pushes back on narratives that place sole responsibility for timeline pressure on the lander providers, saying NASA would still plan a low-Earth-orbit rehearsal analogous to Apollo 9 even if the landers were fully ready today. He describes the rehearsal as including undocking, translation, and re-rendezvous before any lunar attempt. The article notes that NASA has already made a substantial pivot on the suits to protect the overall schedule, and that the agency is embedding personnel and in some cases building hardware itself to keep multiple elements on track.

Isaacman gives examples of that support: if NASA had not deployed resources the way it did, building parts from subcontractors that had not prioritized NASA or had to prioritize higher DX-rating Department of Defense work, the core stage would not have been ready to meet its time frames. Similar support has been required for Orion, including the assignment of thermal protection system specialists. He says everybody needed help, and everybody needed NASA to be NASA and put people in the field and do things. That support has extended to Blue Origin as well, including the use of C-17 aircraft to move large structural domes from the West Coast to the East Coast and the deployment of hundreds of NASA personnel to assist with pad recovery and anomaly investigation work, including welding support. Isaacman says it is incredibly important to get the Blue mission done, and that NASA is prepared to accept limited schedule movement if it yields additional useful data or operational experience. He says that if it means slipping another month to put something else in the Blue Origin lander to get a little more from a team that has never kept people alive in space outside New Shepard, that is something NASA would want. The article concludes with the stated intent to maximize learning on Artemis III and to close the remaining gaps relative to historical lunar landing profiles by demonstrating an uncrewed lander before proceeding to the lunar surface, closing whatever gaps exist between the Apollo 9 mission and Apollo 11.

A third broad group concerns the vulnerabilities, epistemic behavior, and personal philosophy surrounding large language models. How We Took Over ChatGPT Accounts is a Simon Willison's Weblog entry summarizing a report from Parameter Labs. It describes an exploit chain found against ChatGPT's implementation of the artifacts pattern, now called HTML previews, where interactive custom HTML apps are displayed in a sandboxed iframe. The central requirement is that HTML in these previews stay tightly sandboxed to avoid untrusted code interacting with OpenAI APIs and accessing private data. Parameter found that OpenAI uses the MathLive JavaScript library as part of its graph support, and that library includes a virtual keyboard which can be triggered in the parent page via postMessage from the sandboxed iframe. The receiving code uses innerHTML, so anything that sends a message can add custom HTML to the parent page, and there was no mechanism to prevent other sandboxed frames, including HTML preview frames, from calling that method.

This meant untrusted code in an iframe could insert HTML. While innerHTML does not directly execute script elements, it could be used to inject a new iframe without sandbox restrictions. OpenAI's content security policy blocked inline scripts entirely and only allowed script loading from a handful of domains, but one of those domains was YouTube, and the YouTube oEmbed endpoint could be used to reflect JavaScript back in a way that executed it. Once the attack had JavaScript execution on ChatGPT, it was trivial to fetch private user data from APIs and exfiltrate it via an image reference. The end result was that a ChatGPT share URL, when visited by another signed-in user, could exfiltrate their private data or take actions on their behalf. The entry notes that OpenAI closed the hole and paid out a five-thousand-dollar bug bounty, which Simon Willison describes as a bit cheap for a vulnerability this serious.

Are LLMs good or bad Bayesians? is a Marginal REVOLUTION post built around an exchange between Bryan Caplan and ChatGPT. Caplan argues that language models behave as bad Bayesians: when presented with counterevidence, they should move toward or against the presenter with roughly equal probability, depending on whether the argument was better or worse than expected, but in fact they seem either to move toward the speaker or not change. ChatGPT responds by identifying a genuine and important failure mode in AI reasoning, saying that large language models tend to treat a plausible counterargument as evidence in favor of its conclusion rather than comparing the counterargument to what they should have expected to hear. The post asks whether the tendency will be fixed, whether it should be fixed, and whether good prompting can undo it, describing the scope and limits of Bayesian reasoning as truly at stake.

Experiments with Monstrousness is a longer Contraptions essay by Venkatesh Rao, framed as a reintroduction to new readers on Halloween. It opens by noting that more than a thousand new readers signed up over the previous week, mainly because of the post Cognitive Mercantilism, taking the newsletter past twenty-seven thousand subscribers. Rao observes that going viral now means people make and share AI-generated summaries, explainer videos and infographics about a post, including in other languages, and he remarks that argument-logic-flow diagrams have also appeared. He welcomes new readers and says this is a good time for a writerly inner-game reintroduction rather than a basic biographical sketch.

The essay's central motto, adopted about a year ago, is Be Slightly Monstrous. As a doctrinal directive, Rao says it means adopting AI without reservations and letting it freely transform the writer. As a philosophical stance, it is a declaration of resistance to what he calls the Nostalgia for the Human that has overtaken the zeitgeist. The most visible sign of his growing monstrousness is that he freely uses AI in every aspect of his writing, and as of a month ago he no longer segregates or declares it, since AI use is now his default. Fully hand-crafted writing is the exception and is labeled. He writes that AI-assisted output is the default, and that this approach has steadily improved with each post.

Rao says he increasingly operates by a set of Strange New Rules in every aspect of his work and life, all connected to AI in one way or another. Half his work is consulting for the leadership of Tenstorrent, which gives him a ringside view of how powerful the technology is. The other half has to do with protocols. The current essay is a rare handwritten one because the topic is personal; otherwise he always uses AI, and he describes enjoying his centaur-writing more with each post. He writes that even many people who declare that they hate generated writing find themselves enjoying the writing and only discover afterward that it was written with AI. He does not try to disguise the AI voice, and he edits AI text the way he edits other human writers, in a maximally laissez-faire way.

He addresses the belief that he must have fine-tuned a model on his large archive, and says no: it is vanilla ChatGPT, because he saves Claude tokens for coding. He believes he can teach people to write this way, but ninety percent of it is the courage to be slightly monstrous and only ten percent is teachable tips and tricks with prompts and harnesses. He says the main requirement is risking the scorn, contempt and opprobrium of tastefully aghast writerly friends. He notes that his successful experiments with AI writing make some people angrier than the failed ones, and that this has confused and infuriated even some of his oldest readers.

Rao traces the philosophy back through his writing. The motto grew out of his essay The Gramsci Gap, which riffed on the Gramsci-Zizek observation that the old world is dying, the new world is struggling to be born, and it is the time of monsters. That essay in turn came out of Disturbed Realities, which followed his working through the complete works of H. P. Lovecraft, J. G. Ballard, and Benjamin Labatut's When We Cease to Understand the World. The Labatut book, he says, was as formative for the current chapter of his writing life as The Office was for the previous chapter. He is working on a short book pulling these ideas together, to be called New Nature.

The previous chapter was the blog Ribbonfarm, best known for The Gervais Principle, a six-part series about The Office exploring sociopathy, cluelessness and loserdom in organizations. His old motto was Be Slightly Evil. Ribbonfarm is now an archived static site haunted by an AI-powered ghost digital twin called vgr_zirp, and he is building a fuller living-digital-twin bot. The essay that seeded the current chapter is On the Design of Escaped Realities from 2015, which argued that being mildly psychotic was a better idea for the coming decade than being mildly neurotic. That idea led to the newsletter name Contraptions and the bot vgr_gramsci, which he describes as much more psychotic than vgr_zirp but much less neurotic.

He invokes Douglas Adams's remark that if life is going to exist in a universe of this size, the one thing it cannot afford to have is a sense of proportion, and says a dash of psychosis works well for 2026. He argues that since Trump Derangement Syndrome and Woke Derangement Syndrome are past their sell-by dates, most people are choosing between AI Derangement Syndrome and Humanism Derangement Syndrome, and his calibrated psychosis includes a stiff dose of the latter. But he emphasizes that Be Slightly Monstrous is a serious philosophy, one he has bet money and reputation on.

To prove the seriousness, he cites his collaboration with artist Simon Denny on an exhibit at this year's Venice Biennale called Monsters Between Worlds, and the founding of a new magazine called Monstrous Times. He defines monstrousness not in the sense of fantasy but in a science-fictional sense, a Labatutian-Lovecraftian-Ballardian sense: a slightly monstrous human is one transformed through co-evolution with emerging technology in a Gramsci Gap, embodying the strange new rules of that technology. This notion emerged from his other big obsession, protocols. Between 2023 and 2026 he ran the Summer of Protocols program for the Ethereum Foundation, which has evolved into the Protocol Institute, where he is Director of Research, currently unpaid. Monstrous Times also came out of the Summer of Protocols, specifically from an idea called Chiang's Law, after Ted Chiang, that science fiction is about strange rules while fantasy is about special people. Rao says the genre is being called protocol fiction, and one of his own contributions is a Lovecraftian pastiche called The Signal Under Innsmouth.

The essay then asks what it means not to embrace co-evolution with technology, and answers that it means being attached to a historically contingent frozen snapshot of what it means to be human, based on people grappling with AI around 2024. He argues that attempting to preserve humanity while navigating the Gramsci Gap may be the most deadening thing a person can do, while embracing monstrousness may tap into the deepest kinds of liveness available today. He mentions two talks on the idea in the last month, one at the Biennale exhibit and one at the Monstrous Times launch party, and says the idea will be part of New Nature. He also mentions liveness as a pet theme, the focus next year of the book club he runs, while this year the club explores divergence. He closes by hoping the introduction provides orientation in the thought spaces he explores with fellow monsters-in-training, and a sense of where the human fraction of ideas like cognitive mercantilism comes from.

Binance Replenishes 90% of Stolen $6.3B, Pledges They'll Cover It All, Reopen Sunday is a Slashdot article drawing on NBC News, CNBC and The New York Times. It reports that Binance has replenished ninety percent of the six point three billion dollars stolen in a record hack, and has pledged to cover the remaining ten percent. Co-CEO Richard Teng posted that users will not bear the financial impact. The remaining ten percent consists of tokens that are less liquid and may take longer to buy. Binance says it set aside around six point three four billion dollars of company funds internally to cover the losses, using part of its one billion dollar Secure Asset Fund for Users, some of its Futures Insurance Fund, and its corporate treasury.

The FBI blames the breach on a North Korea-sponsored group known as TraderTraitor. Law enforcement observed members converting stolen assets into Ethereum, and the FBI expects further laundering through related technology. After the hack, Binance stopped allowing customers to withdraw funds, but said Saturday it was still on track to reopen withdrawals Sunday. The article notes that the exchange initially said the hack may involve a compromised DevOps engineer's laptop, which allowed an attacker to gain access to a service container and submit unauthorized withdrawal transactions, but co-CEO Richard Teng later said the attack involved a more complex process rather than the misuse of a single engineer's permissions.

The New York Times account describes the response inside Binance. Top executives put together an informal war room on the company's internal messaging system with about one hundred participants, including specialized cryptocurrency tracers and security personnel, then transitioned to a mass conference call with one hundred boxes on screen. Employees took turns muting and unmuting, communicating in a mixture of English and Chinese. Among them was co-chief executive He Yi, who is seven months pregnant and worked through the night from her base in Abu Dhabi, taking a call at three in the morning local time. Because almost all crypto transactions are recorded on a public blockchain, analysts could see funds leaving Binance in real time.

The article places the hack in a longer pattern: North Koreans have targeted crypto companies for years, using the proceeds to fund the country's nuclear program, and have developed increasingly complex strategies to steal and launder crypto, a process that can take months. Chainalysis said the attackers deployed laundering patterns not seen before, conducting thousands of transactions over a short period, probably with the help of automation. At the same time, industry colleagues mobilized to intercept some of the laundered crypto. Tether and Circle announced they had frozen stashes of their own coins seized from Binance, a combined total of more than two billion dollars. Binance is now working to gain access to those funds, but two point four billion dollars in crypto is already beyond any issuer's reach. The article includes President Trump's reaction: he said he had always felt cryptocurrency was a little bit dangerous, but that those involved said they have it pretty much under control, and that he would wait to see.

America's Clocks Are Still Falling Back Tonight Despite Trump's Daylight Saving Push is a Slashdot item based on Axios. It reports that clocks fall back one hour tonight to end daylight saving time, even though many Americans want to end the twice-yearly clock change. Congress remains deadlocked about how to make that happen. The clocks are falling back despite a bipartisan House vote in July to make daylight saving time permanent and Trump's subsequent push for Senate action.

The article says most Americans want to end the twice-yearly clock changes, but they do not necessarily favor Trump's solution. Gallup found that forty-eight percent support permanent standard time for the entire year, compared with twenty-four percent who prefer permanent daylight saving time, which Trump wants. The article notes that in 1974 the United States experimented with year-round daylight saving time but reversed course amid concerns about children going to school in the dark. Experts still debate whether permanent daylight saving time is good for people: supporters argue it would provide more afternoon and evening sunlight, while sleep experts warn that darker mornings and lighter evenings can disrupt sleep cycles and could force millions of children to travel to school in the dark. The American Academy of Sleep Medicine prefers permanent standard time because it better aligns with natural sleep cycles. Hawaii and most of Arizona do not observe daylight saving time and will not change. Eliminating it for the rest of America still faces opposition in the Senate and is not expected to be considered by the end of the year, but the upcoming midterms could change the balance of power in Congress and potentially shift the debate.

4 days left: How TechCrunch Founder Summit helps you make more useful connections to scale is a promotional TechCrunch article. It says that on November 4 at Boston's SoWa Power Station, the event offers a full day packed with meaningful networking opportunities, actionable insights, and connections that could change the trajectory of an idea and startup. It urges readers to register before all prices go up at the door.

The window gathers several recurring tensions: AI adoption is defended as inevitable by Apple, NVIDIA, Xbox, and Venkatesh Rao, while AI security flaws and epistemic weaknesses remain under scrutiny. Data centers are becoming local political issues in the run-up to the midterms, and space policy is moving toward a multi-launch model. A major crypto theft and the perennial daylight saving debate complete the day's reporting.

  1. Judge allows NAACP lawsuit over xAI data center pollution to move forward
  2. 4 days left: How TechCrunch Founder Summit helps you make more useful connections to scale
  3. Apple CEO says the company doesn’t ‘want to ship AI for AI’s sake’
  4. Computer Simulations Suggest Building Blocks of Life Didn't Arrive on Prehistoric Comets
  5. Experiments with Monstrousness
  6. An Italian farmer found some unknown seeds in his attic and planted them
  7. Memdoor: A team of 5 Qwen3.8-27B models beats Fable 5 on LiveCodeBench Hard
  8. Altoids Project
  9. How We Took Over ChatGPT Accounts
  10. Binance Replenishes 90% of Stolen $6.3B, Pledges They'll Cover It All, Reopen Sunday
  11. Brazilian Court Sanctions Lawyers for Prompt Injection in Court Filing
  12. Data centers are now a midterm issue, even in Brooklyn
  13. NVIDIA Updates Rubin GPU Specs as First Production Hardware Arrives
  14. Isaacman outlines Artemis III multi-launch checkout sequence and Blue Origin priority
  15. Are LLMs good or bad Bayesians?
  16. Xbox Manager: AI Already Being Used In Xbox Game Development - And Throughout the Game Industry
  17. America's Clocks Are Still Falling Back Tonight Despite Trump's Daylight Saving Push