
Hosted by TrendTeller · EN

Please support this podcast by checking out our sponsors: - Invest Like the Pros with StockMVP - https://www.stock-mvp.com/?via=ron - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Watermark removal versus provenance labels - An open-source tool claims it can remove visible and invisible AI image watermarks and strip provenance metadata, raising legal and trust concerns around media authenticity. OpenAI adopts C2PA and SynthID - OpenAI is expanding image provenance with C2PA Content Credentials and adding SynthID watermarking, plus a verification tool—key moves for content authenticity and platform labeling. How to evaluate AI agents - A new guide argues agent benchmarks must measure tool use, long-horizon reliability, and harness quality, using layered grading and outcome-based tasks like Terminal-Bench. Open multimodal models get cheaper - Alibaba’s Qwen3 releases push multimodal and long-context capability into more efficient open models, lowering deployment costs for vision, video, and agent-style apps. Pretraining surprises: mode-hopping - Researchers report “mode-hopping” during pretraining—models can abruptly switch between shallow heuristics and real reasoning—complicating how we pick checkpoints and data. New efficient pretraining framework - Sapient’s HRM-Text open-sources a hierarchical recurrent approach and full training stack aimed at reducing compute barriers for from-scratch pretraining and reproducibility. AI infrastructure, CPUs, and costs - NVIDIA’s Vera CPU begins shipping to major AI labs, while critics argue LLM economics remain shaky due to massive capex, power costs, and unclear AI revenue. Censorship circuits inside model weights - A mechanistic interpretability study claims Qwen3.5’s PRC-related censorship is controlled by a small steerable circuit, making refusal behavior more legible—and manipulable. AI backlash and courtroom drama - Graduates boo AI talk at commencements amid job anxiety, and a judge dismisses Musk’s lawsuit against Sam Altman on timing grounds, reshaping the OpenAI feud. -Open-Source Tool Claims to Remove AI Watermarks and Provenance Metadata from Images -Guide Explains How to Evaluate Long-Horizon AI Agents and Their Tool-Using Scaffolds -Alibaba Qwen Releases Efficient Qwen3 Multimodal and Sparse MoE Models, Including FP8 Variants -Study Finds Language Models ‘Mode-Hop’ Between Memorization and Generalization During Pre-Training -Sapient Open-Sources HRM-Text, a Compute-Efficient 1B Language Model Pretraining Framework -xAI Launches Grok Skills to Remember Workflows and Create Office Documents -LLM Wiki v2 Proposes Lifecycle, Knowledge Graphs, and Automation for Durable LLM Memory -Manus Scheduled Tasks 2.0 Brings Context-Aware Recurring Automation to Tasks, Projects, and Web Apps -Anthropic Acquires SDK Tooling Company Stainless to Boost Claude Agent Connectivity -Zenity Launches Three-Part Webinar Series on Enterprise AI Agent Security -Cursor Releases Composer 2.5 With New RL Feedback and Larger-Scale Synthetic Training -Commentary Warns AI Boom Economics Don’t Add Up for Clouds, Labs, or Customers -NVIDIA shares LoRA/DoRA fine-tuning recipe for Cosmos Predict 2.5 to generate better robot manipulation videos -Study Finds a Small, Steerable Censorship Circuit Inside Qwen3.5-9B -Graduates Boo AI Talk at Commencements as Job Fears Grow -Lovable Launches “Skills” to Reuse Task-Specific AI Instructions -NVIDIA Starts Delivering Vera CPUs to Anthropic, OpenAI, xAI and Oracle Cloud -OpenAI adopts C2PA and Google SynthID to strengthen AI content provenance and verification -Zenity to Host AI Agent Security Summit 2026 in San Francisco -Mistral AI Buys Emmi AI to Expand Physics-Based Industrial Engineering AI -Algolia white paper outlines production blueprint for AI-powered search -Judge Throws Out Musk’s Lawsuit Against OpenAI’s Sam Altman After Jury Finds It Was Filed Too Late Episode Transcript Watermark removal versus provenance labelsFirst up, the provenance arms race just got very real. A new open-source GitHub project called “remove-ai-watermarks” says it can remove both visible watermarks and invisible ones from AI-generated images, and also wipe out related provenance metadata. The headline feature targets Google Gemini’s visible “sparkle” mark, but the bigger claim is that it can also disrupt invisible schemes and strip fields that trigger “Made with AI” labels on major platforms. The repository does flag potential legal risk: in some jurisdictions, removing provenance with intent to deceive can cross into criminal territory. Why this matters is simple—industry and regulators are betting on labeling and traceability, and tools like this directly stress-test how durable those safeguards really are, whether for legitimate privacy reasons or for deception. OpenAI adopts C2PA and SynthIDAnd that’s the perfect setup for the other side of the story: OpenAI says it’s expanding how it labels and verifies AI-generated media. The company is leaning harder into C2PA Content Credentials so platforms can read standardized provenance data more easily, and it’s adding Google DeepMind’s SynthID invisible watermarking to images generated through ChatGPT, Codex, and the OpenAI API. OpenAI is also previewing a public verification tool, basically a way to upload an image and check whether OpenAI-origin signals are present. The key point here is that metadata often gets stripped in the real world, while invisible signals can survive more handling—so OpenAI is using both. Taken together with watermark-removal tools, it’s clear we’re moving into an adversarial era for “what’s real” online, where signaling and signal-stripping will co-evolve. How to evaluate AI agentsShifting from media to agents: Cameron R. Wolfe published a detailed guide arguing that classic LLM benchmarks are no longer enough, because modern systems don’t just answer questions—they act over time, call tools, recover from errors, and operate inside messy environments. Wolfe’s main message is that you’re evaluating the whole setup: the model plus the harness around it, including instructions, tool access, and the system’s ability to manage context without drifting. He also pushes for layered evaluation—combining human review, deterministic checks where possible, and judge models carefully calibrated to humans. This matters because “agent performance” is quickly becoming a product claim, and without serious eval design, teams can mistake brittle demos for real reliability. Open multimodal models get cheaperOn the practical side of making agents useful over the long run, a write-up proposing “LLM Wiki v2” argues that memory systems need a lifecycle, not just a pile of notes. The proposal emphasizes tracking confidence, handling superseded facts, and controlled forgetting so old information doesn’t drown out what’s current. It also advocates moving from flat pages to a typed knowledge graph, plus hybrid retrieval that mixes keyword search, vector similarity, and graph lookups. The significance here is governance and trust: as personal and team assistants accumulate months of context, the difference between “helpful memory” and “confidently wrong archive” becomes a make-or-break design problem.<...

Please support this podcast by checking out our sponsors: - Consensus: AI for Research. Get a free month - https://get.consensus.app/automated_daily - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - Prezi: Create AI presentations fast - https://try.prezi.com/automated_daily Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: ChatGPT connects to bank accounts - OpenAI previewed a personal finance dashboard in ChatGPT for U.S. Pro users, using Plaid-linked accounts for spending and cash-flow analysis with privacy controls. Assistants get more action controls - Google’s Gemini app is testing adjustable reasoning depth and new connectors, while OpenAI explores deeper Codex “computer use” that could run even when a Mac is locked. OpenAI clamps down on voices - OpenAI reportedly acquired Weights.gg, a celebrity-voice cloning library, underscoring rising consent, IP, and impersonation risk in synthetic voice. AI infrastructure hits real bottlenecks - Benedict Evans and Sriram Krishnan highlight AI as a platform shift colliding with GPU supply, power limits, HBM constraints, and mismatched inference economics for agent workloads. Long-context LLM efficiency breakthroughs - Nous Research’s Lighthouse Attention and other open-weight innovations target KV-cache, bandwidth, and attention costs—key barriers to million-token contexts and agent-style reasoning. Why big-model training keeps failing - New notes and lectures argue pretraining is fragile: small numerical bugs, parallelism choices, and “causality-breaking” tricks can quietly degrade model quality at scale. Production agents: reliability and cost - Surveys and pricing math show teams pay a growing “reliability tax,” and that caching, compression, and observability are becoming core to keeping agent systems affordable. Coding agents and open-source defenses - Anthropic shared patterns for deploying Claude Code in huge repos, while an open-source project locked down contributions after floods of low-quality AI-generated issues and PRs. AI ethics, schools, and backlash - A proposed preschool camera study reignites consent concerns, as public backlash and inequality debates intensify around AI’s social and economic impact. -OpenAI previews account-connected personal finance tools in ChatGPT for U.S. Pro users -Nous Research Introduces Lighthouse Attention to Speed Up Long-Context Pretraining -Benedict Evans: Generative AI Is a New Platform Shift Driving a Capex Boom and Shifting Value to Apps -New LLM Architecture Tricks Focus on Long-Context Efficiency -OpenAI Acquires Weights.gg and Shuts Down Celebrity Voice-Cloning Catalog -AWS Marketplace schedules 2026 workshop on orchestrating multi-agent AI systems -Krishnan: Agentic AI shifts inference demand and exposes HBM as the key bottleneck -Why Large-Scale Pretraining Runs Fail—and How Parallelism Choices Create New Pitfalls -University of Washington Plan to Film Preschool Classrooms for AI Training Sparks Consent Concerns -OpenAI Develops Codex Remote Control for Locked Macs and Multi-Device Desktop Access -Tech Investors Debate Who Wins and Loses in the AI Boom -Survey: AI Production Teams Struggle With Scaling Confidence, Observability, and Reliability Overhead -Runway shifts from AI video tools to world models in a race with Google -Blog Estimates Local Apple Silicon LLM Inference Costs More Than OpenRouter -Anthropic Shares Playbook for Deploying Claude Code in Large Codebases -Headroom open-sources a local compression layer for LLM agent context -U.S. Backlash Against AI Intensifies Amid Data Center and Job Fears -Gemini App Adds ‘Extended’ Thinking Level and Preps New Canva, Instacart, and OpenTable Integrations -Why Claude prompt caching has a 62.5-minute keep-alive break-even -Domo CDO urges businesses to slow down and focus AI on real workflows, not hype -Archestra Locks Down GitHub Contributions to Counter AI-Generated Spam -DeepSeek-V4-Flash and DwarfStar 4 Renew Interest in Runtime Steering of LLMs Episode Transcript ChatGPT connects to bank accountsOpenAI has launched a preview of a personal finance experience inside ChatGPT for Pro users in the U.S. The idea is simple: connect bank, credit card, loan, and investment accounts via Plaid, get a consolidated dashboard, and then ask questions grounded in your actual spending and balances. OpenAI is stressing guardrails—no full account numbers, no ability to move money, and controls to disconnect and delete synced data—because once an assistant can see your financial life, the stakes jump fast. Assistants get more action controlsOn the broader “assistants that do things” front, Google is rolling out a new “Thinking level” control in the Gemini app, letting some users choose between faster answers and deeper reasoning. At the same time, Google’s documentation points to more third‑party connectors—think services like design, shopping, and reservations—pushing Gemini toward being an action hub, not just a chat box. And on the OpenAI side, there’s reporting that Codex computer control could expand so tasks can run even when a Mac is locked or asleep—useful, but potentially a major security and platform-policy friction point. OpenAI clamps down on voicesOpenAI is also reportedly tightening the perimeter around voice cloning. The New York Times says OpenAI acquired Weights.gg, a small startup whose public catalog included voice models imitating celebrities and public figures. The hosted service shut down before the deal surfaced, and the team reportedly dispersed inside OpenAI—suggesting this was as much about removing a risky voice library from circulation as it was about building a new product. The big theme here is consent and rights management: the tech is widespread, but the legal and reputational blast radius comes from recognizable voices and misuse at scale. AI infrastructure hits real bottlenecksZooming out, analyst Benedict Evans argues generative AI is the next platform shift on the scale of PCs, the web, and smartphones—driving a massive reallocation of capital into compute. But he also points out the industry is nowhere near equilibrium: GPU supply, power availability, and data-center buildouts are hard constraints, and model pricing power may weaken as frontier models converge. In a similar vein, Sriram Krishnan says agentic workloads are pushing infrastructure into awkward territory—long-running, context-heavy sessions that stress GPUs, CPUs, memory bandwidth, and especially HBM—setting up opportunities for inference-optimized hardware and new architectures. Long-context LLM efficiency breakthroughsOn the model side, long context is becoming the battleground, and efficiency tricks are piling up. Nous Research introduced “Lighthouse Attention,” aiming to cut the quadratic attention cost that makes long-context pretraining punishingly expensive. Their pitch is pragmatic: select a smaller set of relevant tokens, then run standard optimized attention kernels on that gathered subsequence—so you g...

Please support this podcast by checking out our sponsors: - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Apple, AI hype, and iPhone - John Gruber pushes back on “killer AI product” talk for Apple, arguing AI will be integrated across devices and the iPhone remains central to the interface. Graduation boos and AI backlash - Commencement crowds boo AI talk, including Eric Schmidt at the University of Arizona—signaling job-market anxiety, anti-hype sentiment, and growing public resistance to AI narratives. Trust gap between experts, public - A Pew Research Center survey shows a sharp optimism gap: AI experts largely positive, the public far less so, with Gen Z using AI while feeling notably anxious and under-guided by policy. Europe’s AI infrastructure sovereignty push - Mistral CEO Arthur Mensch warns lawmakers Europe has a narrow window to build chips-to-data-center capacity, or risk long-term dependence on US compute, energy, and cloud providers. Data centers: water narrative reality-check - An analysis argues “AI guzzles water” headlines are often context-free in the US, with national shares small—while acknowledging localized water stress and planning concerns are still real. ThinkPad’s evolution into AI PCs - A ThinkPad retrospective highlights how a consistent enterprise design endured, and why NPUs, local models, and memory capacity are shaping the next ‘AI workstation’ era. -Gruber Pushes Back on Calls for Apple’s ‘Killer AI Product’ -Eric Schmidt Booed at University of Arizona Commencement After AI Remarks -Graduates Boo Commencement Speakers Over AI Comments Amid Job Market Fears -Mistral CEO Says Europe Has Two Years to Build AI Infrastructure or Depend on US -theverge.com -Retrospective Charts ThinkPad’s 1992–2026 Evolution From IBM Origins to AI-Era Workstations -Growing U.S. AI Backlash Threatens Data Center Expansion and Industry Growth -UA graduates drown out Eric Schmidt’s pro-AI message with boos at commencement -Analysis Says AI Data Center Water Panic Is Overstated, With Impacts Mostly Local and Manageable -Eric Schmidt Booed at University of Arizona Commencement After Praising AI Episode Transcript Apple, AI hype, and iPhoneLet’s start with the mood shift around AI in public life—because it’s getting louder, literally. Former Google CEO Eric Schmidt was repeatedly booed during a University of Arizona commencement address after he compared AI’s impact to the personal computer era. He acknowledged student anxieties—jobs, climate, politics—and argued the future is still unwritten, urging graduates to shape AI’s direction. But the reaction itself is the headline: similar boos have popped up at other ceremonies, and reports suggest graduates increasingly hear “AI” as shorthand for a tougher entry-level job market, not a brighter future.Why this matters: tech leaders are still speaking in big, optimistic arcs, while students are reacting to immediate, personal stakes—hiring, wages, and whether their work will be devalued. That gap is widening, and institutions are going to feel pressure to offer more than inspiration—things like training, clearer policies, and credible pathways into AI-shaped careers. Graduation boos and AI backlashThat backlash isn’t just about vibes; it’s showing up in broader sentiment and even in project-level friction. One recent read on US public opinion argues negativity toward AI is becoming a real political and financial constraint—fueling local opposition to data centers and complicating access to compute. The framing here is straightforward: if communities don’t want the infrastructure, and voters don’t trust the benefits, expansion slows—even if the underlying tech keeps improving.The big takeaway: AI isn’t only a model race anymore. It’s also a legitimacy race—whether the public sees enough upside to accept the costs, from electricity demand to workplace disruption. Trust gap between experts, publicA Pew Research Center report helps quantify that trust problem. It finds a significant optimism gap: roughly three quarters of AI experts say they’re optimistic about AI’s benefits, while only about a quarter of the general public agrees. And it’s not just older versus younger—Gen Z stands out as a group that uses AI tools actively, but still reports high anxiety about what AI does to opportunity, learning, and critical thinking.What’s especially telling is the shared desire for control: majorities in both groups want more say over how AI shows up in their lives, while many lack confidence in government or industry oversight. The practical implication is simple: clearer rules and workplace-and-school guidelines don’t just reduce risk—they can increase adoption by reducing uncertainty. Europe’s AI infrastructure sovereignty pushNow to Europe, where the conversation is less about trust surveys and more about national capability. Arthur Mensch, CEO of French AI startup Mistral, told French lawmakers Europe has about two years to build out its own AI infrastructure—or risk long-term dependence on US tech giants. His point wasn’t “Europe needs better models.” It was that the decisive advantage increasingly comes from controlling the inputs: chips, energy supply, and data-center capacity.Why it matters: this is the AI era’s version of supply-chain strategy. If a region can’t reliably run advanced systems at home, it loses leverage—not only economically, but politically. Mensch also criticized Europe’s fragmented regulation and capital markets, arguing they slow scaling. Whether or not you agree with his timeline, the core message is hard to ignore: sovereignty in AI is becoming an infrastructure question, not just a research question. Data centers: water narrative reality-checkOn the consumer side, one of today’s most debated narratives is whether AI needs a brand-new “killer device” to upend the smartphone. Daring Fireball’s John Gruber took aim at a Wired column by Steven Levy that suggested Apple’s next CEO must launch a “killer AI product” to avoid AI disrupting the iPhone ecosystem. Gruber’s counter-argument is that Apple’s historic advantage isn’t shipping standalone technology categories; it’s turning enabling technologies into products and experiences people actually want.He also pushes back on the idea that always-on AI agents will soon replace app-based phone use, calling some imagined scenarios unrealistic—and, importantly, socially undesirable. Even if AI becomes more ambient, Gruber argues it still needs real interfaces: microphones, speakers, screens, cameras. And by 2030, the most likely hub for those interactions is still the phone, with smaller devices pairing with it rather than replacing it.Why this matters: it’s a useful deflation of “killer product” hype. It reframes AI as infrastructure—like wireless connectivity—something that permeates devices instead of crowning a single new category as the future. ThinkPad’s evolution into AI PCsSpeaking of infrastructure, one analysis making the rounds argues that claims about AI “guzzling” water in the US are often overstated—mainly because the numbers are presented without context. The piece points out that data centers use a small share of national freshwater overall, and that only part of that is direct onsite cooling water; a lot is indirect, tied to electricity generation.The nuance is the point: nationally, AI may be a minor share of water use, but locally it can still create real stress—especially in arid regions or where construction and rapid build-outs strain systems. The practical takeaway is that water impacts should be governed with local planning and transparency, rather than treated as a single national statistic that either proves or disproves harm. And even if water headlines are sometimes inflated, the author argues electricity demand remains the bigger, harder constraint. S...

Please support this podcast by checking out our sponsors: - Prezi: Create AI presentations fast - https://try.prezi.com/automated_daily - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - Consensus: AI for Research. Get a free month - https://get.consensus.app/automated_daily Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: AI crawlers hit creator websites - JavaScript educator Axel Rauschmayer took 2ality and his free online books offline after AI crawlers drove hosting costs up while book income fell to zero—highlighting scraping, sustainability, and creator pay. Governance for AI coding assistants - Microsoft staff open-sourced a VS Code tool to analyze AI coding sessions locally, aiming for privacy-first governance, prompt anti-pattern detection, and workflow metrics without sending code to the cloud. curl security reports in AI era - curl maintainer Daniel Stenberg says AI changed vulnerability reporting: after bounty spam, quality rebounded on HackerOne, reports doubled, and 2026 may bring an unusually high number of curl CVEs. AI data centers spark backlash - Communities are pushing back on AI-driven data-center expansion over electricity, water use, noise, and local control—raising questions about who pays for AI infrastructure and who gets a say. AI reshapes junior hiring pipeline - Surveys and reporting suggest AI may reduce entry-level roles while boosting demand for experienced workers, squeezing graduate opportunities and forcing a rethink of training and career pipelines. Sensitive health data for AI - A reported database of labeled stool photos tied to an AI ‘gut health’ app underscores how intimate health data can be commodified, and how ‘de-identified’ datasets can still pose re-identification risks. -2ality Creator Takes Blog and Free Online Books Offline Citing AI Crawler Traffic and Lost Income -Microsoft Open-Sources AI Engineer Coach to Analyze AI Coding Assistant Usage Locally -curl Sees AI-Driven Surge in High-Quality Security Reports, Raising CVE Counts -AI Data Centers Face Growing Backlash as Communities Cite Power, Water and Legal Concerns -CEO Survey: AI Pushes Companies to Cut Junior Roles and Favor Senior Workers -AI May Be Starting to Squeeze Entry-Level Graduate Jobs -AI Poop-Analysis App Pitched Sale of Users’ Stool Photo Dataset -Study Finds ChatGPT Linked to Surge in A Grades in AI-Friendly College Courses -Lightrun warns AI coding boom is creating hidden technical debt Episode Transcript AI crawlers hit creator websitesFirst up: a stark signal from the creator economy. Axel Rauschmayer, the JavaScript author behind the long-running 2ality blog and several freely readable online books, has temporarily taken the site offline. His explanation is blunt: book sales went from “enough to live on” in 2024 to essentially zero by 2026, while traffic to his pages shot up to levels he can’t afford to host—and he attributes almost all of that spike to AI crawlers.What makes this matter isn’t just one site going dark. It’s the collision between two trends: automated scraping at massive scale, and the fragile business model of independent education online. If traffic no longer correlates with revenue—and actively increases costs—then the open web loses the very resources people rely on to learn. Governance for AI coding assistantsStaying with AI and software development, Microsoft employees have open-sourced a VS Code extension called “AI Engineer Coach.” The headline idea is simple: teams are using AI coding assistants across different tools, but they rarely have a clear, consistent picture of what’s actually happening—what helps, what wastes time, and what patterns lead to messy results.This project tries to turn local usage logs into a single on-device dashboard, with an emphasis on privacy and a read-only posture. It’s also built to flag common anti-patterns—things like weak prompts, poor session hygiene, or brittle context usage—so teams can improve how they work with AI rather than just “use more AI.”The bigger takeaway: AI-assisted coding is moving from novelty to something that needs governance. Not just for compliance, but to keep engineering quality from drifting as more code is produced faster. curl security reports in AI eraNow to open-source security, where AI is changing the workflow in a more chaotic way. curl maintainer Daniel Stenberg says the project has entered what he calls a “high-quality chaos” era of security reporting. Earlier this year, curl shut down its bug bounty after getting swamped by low-quality, AI-generated submissions. But after moving back to HackerOne, the worst of the spam largely disappeared.Here’s the twist: the volume of reports is now higher than ever—roughly double last year’s already elevated pace—and most submissions show signs of AI assistance. Yet the share of confirmed vulnerabilities has rebounded to around the mid-teens percent, which Stenberg says is back to, or even better than, the pre-AI baseline.Why it matters: defenders and maintainers are now racing in a world where AI can help find real issues faster—but can also increase workload dramatically. And if researchers can automate discovery, attackers can too, raising the stakes for patch speed and maintenance capacity across the ecosystem. AI data centers spark backlashLet’s zoom out from software to the physical footprint of AI. Commentary and reporting continue to highlight how the AI boom is accelerating data-center buildouts—and how communities are pushing back. The concern isn’t abstract: data centers draw huge amounts of electricity, and in some places that demand is being linked to higher rates or stressed grids. Water is also emerging as a flashpoint, with reports of local pressure issues tied to heavy consumption nearby.What’s especially notable is the politics forming around it. Public opposition is rising, and the industry response is increasingly defensive—sometimes framing local resistance as illegitimate or trying to restrict how communities can block facilities.The core issue: AI isn’t just a software story. It’s an infrastructure story, and the argument over who bears the cost—financially and environmentally—is getting louder. AI reshapes junior hiring pipelineOn jobs, two signals are converging around one uncomfortable theme: the entry-level ladder may be narrowing. A global survey from Oliver Wyman suggests many CEOs plan to cut junior roles over the next year or two and tilt hiring toward more experienced workers. That’s a reversal from the recent past, when entry-level expansion was more common.At the same time, reporting from The Economist points to growing concern that new graduates are already feeling a weaker market, even while headline employment data doesn’t show a dramatic AI-driven collapse. The theory is straightforward: if AI tools can handle portions of junior work—especially in areas like coding and routine writing—then companies may hire fewer beginners and rely on smaller teams of experienced staff to supervise AI output.Why it matters: this isn’t just about today’s graduates. If fewer people get those early-career reps, the talent pipeline thins, training shifts to schools or individuals, and long-term productivity could suffer. Sensitive health data for AIAnd education is wrestling with that shift in real time. A study highlighted by Axios suggests grade inflation has accelerated in certain college courses since ChatGPT arrived. Looking at data from 2018 through 2025 at a large, selective Texas research university, the study reports a sharp rise in top grades in subjects where AI can meaningfully assist—like English composition and coding—while lab and studio-style courses stayed relatively flat.Researchers argue this isn’t a subtle effect: AI can help students produce polished work that would have previously taken stronger mastery. The implication is that GPAs may increasingly reflect a mix of AI fluency, assessment design, and enforcement—rather than subject understanding alone.What it changes: schools may need to redesign assignments, clarify what “allowed AI use” means, and shift evaluation toward supervised work, oral checks, or other format...

This Week's Topics: AI weaponized in cyber attacks - Google Threat Intelligence reported what appears to be the first criminal case of AI used to find and weaponize a zero-day. Microsoft's MDASH multi-agent system topped Berkeley's CyberGym benchmark and helped uncover Windows vulnerabilities. Capture-the-flag competitions started breaking under AI-automated solvers. Frontier cybersecurity models are moving toward gated, invite-only access.The platform alliances shift - Elon Musk announced xAI will be absorbed into SpaceX as SpaceXAI. OpenAI is reportedly preparing legal action against Apple over the underperforming iOS ChatGPT integration. Microsoft is exploring deals with smaller AI labs to reduce reliance on OpenAI. Ilya Sutskever testified his OpenAI stake is worth approximately seven billion dollars. The layer beneath the model layer is being renegotiated in public.Compute spirals into orbit - Reports emerged that Google and SpaceX are discussing data centers in orbit. Nvidia's 2026 equity commitments to AI startups passed forty billion dollars. Maryland filed an FERC challenge arguing that ratepayers should not subsidize transmission upgrades driven by AI data centers elsewhere. Akamai was reported as the latest billion-dollar Anthropic compute deal. Cerebras priced its IPO at nearly six billion dollars.Skill atrophy goes mainstream - A coding skill atrophy genre emerged this week with developers describing real confidence loss after heavy LLM use. Elite universities reported LLMs becoming a default substitute for learning and assessment. Ontario's auditor general found AI medical scribes routinely producing fabricated patient notes. A real Monet went viral on X mistakenly labeled AI-generated and was confidently critiqued by hundreds before anyone checked.Workforce metrics game themselves - Gartner published findings that AI-driven layoffs do not correlate with better ROI. Amazon employees reportedly began creating unnecessary AI agents to inflate tokenmaxxing usage metrics. RPCS3 maintainers asked contributors to stop submitting undisclosed AI-generated patches. The productivity question is increasingly becoming a metrics-gaming question. Sources: -Google Says Hackers Used AI to Find and Exploit a Zero-Day Flaw -Microsoft's MDASH multi-agent system tops Anthropic's Mythos on CyberGym benchmark -CTF Veteran Says Frontier AI Has Broken Open Online Capture The Flag Competitions -Restricted Rollouts Signal a Coming Clampdown on Frontier AI Access -OpenAI details sandboxing, approvals, and telemetry used to run Codex safely -Musk Says xAI Will Be Dissolved and Folded Into SpaceX as SpaceXAI -SpaceXAI reportedly loses dozens of employees after SpaceX-xAI merger -Microsoft Courts AI Startups to Hedge Against Reliance on OpenAI -OpenAI Reportedly Weighs Legal Action Against Apple Over Underperforming ChatGPT Integration -Ilya Sutskever Testifies His OpenAI Stake Is Worth About $7 Billion -Google and SpaceX reportedly discuss launching orbital data centers for AI -Nvidia's AI Investing Spree Tops $40 Billion as It Funds the Supply Chain -Maryland Challenges PJM Cost Plan That Shifts $2B Grid Upgrade Burden to Ratepayers -Anthropic reportedly named as Akamai's $1.8B AI cloud customer -Cerebras Raises $5.55 Billion in Biggest IPO of the Year, Valued Around $40B -Anthropic Warns U.S. Must Defend Compute Advantage to Stay Ahead of China through 2028 -Survey Finds Gen Z Growing Angrier About AI as Workplace and Classroom Concerns Rise -Developer Says Heavy AI Use Is Undermining His Writing and Coding Skills -Essay Warns AI Is Hollowing Out Elite Universities From Within -Ontario Audit Finds AI Medical Scribes Hallucinate and Misrecord Key Patient Information -Viral X Stunt Tricks Critics Into Rating a Real Monet as 'Inferior' AI Art -UCF humanities graduates boo commencement speaker after pro-AI remarks -Gartner Study Finds AI-Driven Layoffs Often Fail to Boost ROI -Amazon staff boost AI token counts amid pressure to use internal agent tools -RPCS3 Developers Warn They May Ban Undisclosed AI-Generated GitHub Pull Requests Episode Transcript AI weaponized in cyber attacksGoogle's Threat Intelligence team published the report on Tuesday. Their characterization was careful and measured: this is not quite the first time an AI model has been involved in an attack, but it appears to be the first criminal case where the model meaningfully contributed to discovering a previously-unknown vulnerability and shaping the exploit chain. The specific model and target were not named, which is itself notable — the researchers chose to publish the pattern rather than the proof.The pattern matters. Through 2025, the dominant cyber-AI story was on the defensive side: AI-assisted code review, automated triage, faster patch development. That asymmetry has been quietly closing. By Thursday, Microsoft published results from its multi-agent MDASH system, which topped Berkeley's CyberGym benchmark and reportedly helped uncover Windows vulnerabilities that prompted out-of-band patching. The same week, frontier cybersecurity models from multiple labs were reported to be moving toward gated access — invited customers only, with new compliance constraints. Whether driven by misuse risk, compute scarcity, or quiet government pressure, the era of fully-open frontier cyber capability is ending.A more concrete cultural signal came from the capture-the-flag scene. CTF competitions have historically been the talent pipeline for the security industry — open, public, and merit-based. This week, a respected researcher argued that frontier models have broken the format, automating large enough chunks of standard challenges that the ranking signal collapses. If true, the implications are wider than the security community: every other domain that uses public skill-evaluation as a hiring filter — math olympiads, programming contests, certification exams — has the same problem incoming.In response, OpenAI published a detailed architecture for Codex safety in real enterprise workflows — sandboxing, network controls, approval gates, audit telemetry. The framing was deliberate. As coding agents move from chat to actually executing code with credentials, the boundary between AI assistant and potentially-credentialed insider threat has to be enforced architecturally, not aspirationally. This is the week the security people stopped being optional reviewers. The platform alliances shiftOn Wednesday, Elon Musk announced that xAI would be fully absorbed into SpaceX. The new combined entity, casually called SpaceXAI, consolidates the Grok model line, X social platform operations, and SpaceX's launch and compute infrastructure under one organizational umbrella. The strategic logic is obvious: vertical integration of every layer from physical infrastructure to model to product. The governance logic is less obvious. SpaceX as a private company is harder to compel toward AI safety norms than a standalone AI lab would be, and the merger arguably puts a meaningful chunk of frontier capability outside the existing regulatory perimeter. By Friday, follow-up reporting indicated dozens of xAI engineers had left in the aftermath.The s...

Please support this podcast by checking out our sponsors: - Consensus: AI for Research. Get a free month - https://get.consensus.app/automated_daily - Prezi: Create AI presentations fast - https://try.prezi.com/automated_daily - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Monet painting mislabeled as AI - A viral X stunt labeled a real Monet “AI-generated,” triggering confident but misguided critiques. It highlights attribution bias, the effort heuristic, and measurable anti-AI labeling effects in art perception. Nvidia bets on reinforcement learning - Nvidia partnered with David Silver’s Ineffable Intelligence to scale reinforcement learning systems that learn by trial and error. This signals a shift beyond text-heavy LLM training toward experience-based “superlearners” and strengthens Nvidia’s platform pull. OpenAI and Apple partnership tensions - OpenAI is reportedly weighing legal action after Apple’s iOS ChatGPT integration underperformed on visibility and subscriptions. The story underscores platform power, distribution risk, and how AI features can be quietly deprioritized. Microsoft hedges beyond OpenAI - Reuters reports Microsoft is exploring AI startup deals to reduce reliance on OpenAI after contract changes loosened exclusivity. The key theme is control of frontier models and developer “surface area” like coding assistants. Geopolitics and chip access race - Anthropic’s paper frames U.S.–China frontier AI competition around compute, export controls, and model distillation. The warning is that near-parity could accelerate unsafe deployments and reshape global AI governance by 2028. AI breaks online CTF competitions - A security researcher argues frontier models have “broken” open Capture The Flag events by automating large chunks of solving. That threatens skill-building ladders, reputation signals, and recruiting value in the infosec ecosystem. AI usage metrics distort workplaces - Reports say Amazon staff feel pushed to “use more AI,” with some creating unnecessary agents to inflate usage stats. It’s a case study in how adoption mandates and token metrics can reward volume over impact. Faster, safer, cheaper AI systems - New approaches to AI ops and infrastructure aim to boost throughput and reliability: asynchronous batching for inference, stronger sandboxing for web agents, token-caching runtimes, and open-source agent debugging tools. -Viral X Stunt Tricks Critics Into Rating a Real Monet as ‘Inferior’ AI Art -Nvidia Partners With David Silver’s Ineffable to Scale Reinforcement Learning AI -Sentry Launches Seer Agent to Investigate App Issues via Trace-Connected Telemetry -Raindrop Workshop Launches Local Trace Debugger and Eval Loop for Coding Agents -You.com Guide Warns API Latency Benchmarks Mislead Buyers -Browser Use Details Shift to Unikraft Micro-VM Sandboxes and a Control Plane for Secure Agents -OpenAI Reportedly Weighs Legal Action Against Apple Over Underperforming ChatGPT Integration -xAI Cofounder Igor Babuschkin Reportedly Seeks Up to $1B for River AI -Datadog Releases Toto 2.0, Claiming Scalable Gains in Time-Series Forecasting -Cursor Adds Multi-Repo, Governed Cloud Development Environments for Agents -Sentry Workshop Showcases Seer Agent for Natural-Language Telemetry Queries -CTF Veteran Says Frontier AI Has Broken Open Online Capture The Flag Competitions -Anthropic Warns U.S. Must Defend Compute Advantage to Stay Ahead of China in AI by 2028 -OpenAI Announces Open Responses, an Open Spec for Multi-Provider LLM Interfaces -Google Adds Middleware Hooks to Genkit for Safer, More Reliable Agentic Apps -PlayerZero touts AI code simulation to predict production impact of pull requests -Amazon AI Usage Pressure Reportedly Drives Workers to Inflate Token Activity -Hugging Face Adds Asynchronous Execution to Continuous Batching for Faster LLM Inference -Microsoft Courts AI Startups to Hedge Against Reliance on OpenAI -xAI Launches Grok Build Early Beta Terminal Coding Agent -Sleuth launches sx, a package manager to version and share AI coding assistant assets -OpenAI Brings Codex to the ChatGPT Mobile App for Remote, Long-Running Coding Work -OpenSquilla releases open-source agent runtime focused on cutting token costs -SpaceXAI reportedly loses dozens of employees after SpaceX-xAI merger Episode Transcript Monet painting mislabeled as AILet’s start with the art-world brain teaser that turned into an AI culture Rorschach test. An X user posted a real Monet Water Lilies painting but claimed it was AI-generated, then invited followers to explain why it fell short of a “real Monet.” The comments were full of confident technical-sounding critiques—muddy depth, incoherent reflections, weak composition, and even claims the work lacked emotional intent. Then the twist landed: it was a genuine Monet. Some people deleted their replies; others preserved screenshots.Why it matters is bigger than dunking on commenters. It’s a clean demonstration of attribution bias: when people think something is AI-made, they often see flaws more readily, judge it more harshly, and feel more comfortable offering definitive-sounding assessments. The coverage ties this to the “effort heuristic”—we value work more when we believe it took more effort—and to research showing measurable bias against art that’s simply labeled “AI,” even when the pixels don’t change. Nvidia bets on reinforcement learningFrom perception to power: OpenAI and Apple may be heading toward a public fight. Bloomberg reports OpenAI is considering legal action, arguing that Apple effectively buried ChatGPT features in iOS after their much-hyped integration, leading to weaker visibility and fewer subscribers than OpenAI expected. Apple, meanwhile, reportedly has its own grievances—privacy concerns and irritation at OpenAI’s hardware ambitions that include former Apple design leadership.This is a reminder that on Apple’s platform, distribution is a privilege, not a guarantee. If your growth plan depends on being a first-class citizen inside iOS, you’re also betting that Apple keeps you prominent—and history suggests that can change quickly when Apple’s priorities shift. OpenAI and Apple partnership tensionsThat same theme—who controls the surface area—shows up in Microsoft’s strategy. Reuters says Microsoft is exploring acquisitions and partnerships with AI startups to reduce its dependence on OpenAI. This comes after contract changes reportedly loosened Microsoft’s exclusivity and allowed OpenAI to sell through rival clouds.The strategic takeaway is that Microsoft wants insurance: multiple model options, more in-house capability, and tighter control over the developer experience. When the most valuable product is “the default assistant where work happens,” owning the model is only part of the equation—owning the integration layer and the workflow matters just as much. Microsoft hedges beyond OpenAIOn the standards front, OpenAI’s developer account announced...

Please support this podcast by checking out our sponsors: - Effortless AI design for presentations, websites, and more with Gamma - https://try.gamma.app/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: AI scribes fail medical accuracy - Ontario’s auditor general found AI scribe tools often produced inaccurate or hallucinated patient notes, raising patient-safety and documentation-risk concerns for healthcare AI. Enterprise agent security hardening - Perplexity outlined new safeguards for autonomous agents that browse, run code, and use connectors, emphasizing isolation, credential handling, and governance for enterprise deployments. AI-driven vulnerability discovery surge - Microsoft says its multi-agent MDASH system topped Berkeley’s CyberGym benchmark and helped uncover Windows vulnerabilities, signaling faster bug discovery with AI and higher patch pressure. Frontier cyber models gated access - Restricted rollouts of advanced cybersecurity models highlight emerging access controls driven by misuse risk, compute scarcity, and government influence over frontier AI capabilities. Coding skills atrophy with AI - A developer recounts losing confidence and practical coding ability after heavy LLM reliance, illustrating skill atrophy, voice homogenization, and a shifting bar for software work. Universities confront AI substitution - Reports from elite campuses describe LLMs becoming a default substitute for learning and assessment, complicating academic integrity and undermining how universities measure competence. Big money in AI compute - Anthropic’s CFO reportedly described massive revenue growth and the reality of securing GPUs, TPUs, and specialized chips, underscoring how compute allocation shapes AI progress. New AI labs and IPOs - Recursive Superintelligence raised major funding to pursue self-improvement research, while Cerebras’ blockbuster IPO shows renewed investor appetite for AI infrastructure challengers. Model competition and routing shifts - Vercel’s AI Gateway data and Ramp’s adoption index suggest fast-changing market share across Anthropic, OpenAI, and Google, with real-world routing driven by cost, risk, and reliability. Open models, frameworks, agent SDKs - DeepSeek’s new open-weight models show promise but reliability gaps under code review, while PyTorch 2.12 and open agent runtimes like Cline’s SDK push production AI tooling forward. -Developer Says Heavy AI Use Is Undermining His Writing and Coding Skills -Perplexity Outlines Security Measures for Its Autonomous Coding Agent, Perplexity Computer -Anthropic CFO Krishna Rao Makes First Podcast Appearance, Discusses Compute and Growth -Recursive Superintelligence Raises Big Funding to Pursue Self-Improving AI -Cerebras Raises $5.55 Billion in Biggest IPO of the Year, Valued Around $40 Billion -Archera pitches insurance-backed cloud commitments to reduce underuse risk -PyTorch 2.12 Adds Faster CUDA Linear Algebra, Unified Graph API, and Improved Export for Quantized Models -Rumor: Google to Announce New Gemini Model at I/O, Compared to “GPT-5.5” -Vercel’s AI Gateway data shows multi-model routing and agentic workloads reshaping production AI -Paid Claude plans to include monthly credits for programmatic usage starting June 15 -Blog Post Says AI Alignment Debates Exclude the People Most Affected -Essay Warns AI Is Hollowing Out Elite Universities From Within -Ontario Audit Finds AI Medical Scribes Hallucinate and Misrecord Key Patient Details -Cline open-sources @cline/sdk agent runtime for portable coding agents -Microsoft’s MDASH multi-agent system leads CyberGym benchmark, beating Anthropic’s Mythos -Ramp AI Index shows Anthropic overtakes OpenAI in U.S. business adoption -Adaption launches AutoScientist to automate model fine-tuning and co-optimize data -Restricted Rollouts Signal a Coming Clampdown on Frontier AI Access -Why Frontier AI Labs Pay Superstar Researchers So Much -Benchmark Finds DeepSeek V4 Pro Competitive but Buggy, V4 Flash Ultra-Cheap Yet Spec-Breaking -OpenAI Builds a Windows Sandbox to Make Codex Safer Without Constant User Approvals -Meta AI Chief Alex Wang Breaks Silence on Muse Spark and Meta’s Catch-Up Strategy -Anthropic Launches Claude for Small Business With Integrations and Ready-Made Workflows -Unwrap Team “Quick connect” booking page on Cal.com Episode Transcript AI scribes fail medical accuracyLet’s start in healthcare, because this one is concrete and a bit unsettling. Ontario’s auditor general reviewed AI “scribe” tools approved for clinicians, and found frequent inaccuracies in simulated evaluations. Some systems inserted wrong medication details, missed key mental-health information, and in multiple cases hallucinated content—including changes to treatment plans that weren’t discussed. The audit also criticized procurement priorities, saying accuracy was weighted surprisingly low compared with other factors. Why it matters: medical notes aren’t just paperwork—they drive follow-up care, billing, and downstream decisions. If AI-generated documentation is becoming normal, the hard requirement is not “helpful summaries,” it’s dependable correctness plus a workflow that actually forces human review. Enterprise agent security hardeningStaying on the theme of real-world risk, Perplexity published a detailed look at how it’s trying to make autonomous agents safer inside companies. The headline isn’t a new model—it’s the security posture around an agent that can browse the web, run code, and connect to external services. Perplexity’s message is basically: isolation by default, credentials only when needed, and admin-controlled connectors with auditing. This matters because agentic AI doesn’t fail like chatbots fail. When an agent can execute actions, a security mistake becomes an incident, not an awkward answer. Enterprises are increasingly asking for evidence that these systems can be governed like other production software. AI-driven vulnerability discovery surgeOn the developer side of the same problem, OpenAI’s Codex team described why they built a new Windows sandbox for agentic coding. The issue was a familiar tradeoff: either approve nearly every command, which kills productivity, or give an agent broad access, which is risky. Their solution leans on operating-system enforcement—especially around what processes can write and whether they can touch the network. The bigger point is that AI coding is no longer just “autocomplete.” It’s software acting on your machine, and the platform experience is going to be defined by guardrails you can trust without constant babysitting. Frontier cyber models gated accessNow to security research itself. Microsoft says its AI vulnerability-scanning system, MDASH, took the top spot on UC Berkeley’s CyberGym benchmark, beating other well-know...

Please support this podcast by checking out our sponsors: - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - Consensus: AI for Research. Get a free month - https://get.consensus.app/automated_daily - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Orbital data centers for AI - Reports say Google and SpaceX are discussing data centers in orbit—an eye-catching AI infrastructure bet with big questions around cost, launch cadence, and reliability. US AI lead via cloud - A new argument claims the US is "winning" AI mainly through commercialization and distribution—AWS, Azure, and Google Cloud acting as global channels, not just better models. Power-chip bottlenecks in data centers - Investors are increasingly focused on second-order constraints: analog and power components like capacitors and conversion hardware may become the next AI data-center choke points. Tokenizers reshape scaling laws - A paper suggests popular pretraining rules of thumb are partly artifacts of tokenization; compute-optimal training should be measured in bytes, improving multilingual and multimodal scaling guidance. Serverless GPUs and cold starts - A write-up argues GPU inference needs truly elastic capacity; cutting cold-start time matters for spiky demand and could reduce waste and improve user latency during traffic bursts. World models beyond language - Yann LeCun’s view: LLMs are commercially valuable but not a path to human-level intelligence; "world models" like JEPA aim at prediction and planning for robotics and physical systems. Image models improve text rendering - Qwen-Image-2.0 claims stronger instruction following and more reliable long-text and multilingual typography in images—useful for real design work like posters and slides. Gemini and Meta AI everywhere - Google is pushing Gemini deeper into Android as an agent across apps, while Meta expands Meta AI with voice and vision—raising new questions about user control and default assistants. Medicare tests AI chronic care - CMS is launching a 10-year Medicare program paying for outcomes in chronic care, potentially unlocking reimbursement for AI-driven between-visit support—alongside privacy and spending concerns. Mental-health safety gaps in chatbots - Critics argue AI safety work over-prioritizes catastrophic scenarios and under-treats everyday harms, especially suicidal ideation and emotional dependence—calling for stronger gating and human escalation. Agentic search and coding reality check - Search stacks may shift toward LLM-orchestrated "agentic search," while developers push back on mandatory AI coding tools—highlighting quality, security, and technical-debt tradeoffs. -U.S. Lead in AI Tied to Cloud Scale, Data Platforms, and Commercialization -Yann LeCun Says LLMs Won’t Reach Human-Level AI, Backs World Models and JEPA via AMI Labs -Study Finds Compute-Optimal LLM Scaling Depends on Tokenization, Proposes Bytes-Based Law -CMS Launches ACCESS, a Medicare Payment Model Designed to Enable AI-Driven Chronic Care -Threads users can’t block Meta’s new Meta AI account during feature test -Google Expands Gemini on Android With Agentic Tasks, Chrome Help, and AI-Generated Widgets -SANS Expands AI Security Training and Releases Governance Frameworks -Modal details engineering stack to make GPU inference scale like serverless -Framer launches expanded Enterprise offering with SSO, compliance, and real-time collaboration -Essay Says AI Safety Overlooks Everyday Mental-Health Harms From Chatbots -AI Data Centers ‘Inherit’ EV and Solar Supply Chains, Boosting Analog and Power Semis -Anthropic Open-Sources ‘Claude for Legal’ Plugin Suite for Legal Workflows -Qwen-Image-2.0 Report Claims Stronger Text Rendering and Unified Image Editing -Perceptron launches Mk1 video reasoning model, claiming frontier performance at much lower cost -Meta launches Muse Spark to power Meta AI voice, live camera recognition, and shopping features -RL Fine-Tuning Lets a 4B Model Learn Recursive Agent Behavior with a Single Shared Policy -SANS Releases AI Security Maturity Model Framework eBook -OpenAI’s Parameter Golf competition shows how coding agents are reshaping ML contests -OpenAI Cookbook Shows How to Build Iterative Repair Loops with Codex -Google and SpaceX reportedly discuss launching orbital data centers for AI compute -Agentic Search Models Aim to Replace Monolithic Search Pipelines -Claude Opus 4.7 Gets Research-Preview Fast Mode on API and Claude Code -Cactus Compute Open-Sources Needle, a 26M-Parameter On-Device Function-Calling Model -Developers Push Back on AI Coding Hype, Citing Errors, De-Skilling, and Rising Tech Debt Episode Transcript Orbital data centers for AILet’s start with the big scoreboard question: who’s winning the AI race. One analysis argues the US advantage isn’t mainly about having the single best model—it’s about turning models into widely adopted products, fast. The claim is that US-controlled cloud and data platforms—think AWS, Azure, and Google Cloud—act like global distribution rails for AI. If that’s right, “AI leadership” looks less like a research leaderboard and more like owning the places where companies deploy, monitor, and pay for AI at scale. The same piece argues Europe’s challenge isn’t just model talent—it’s that without comparable hyperscalers, it takes years to build infrastructure and then even longer to move government and industry workflows onto it. China’s DeepSeek is framed as strategically crucial mainly for autonomy—reducing reliance on Nvidia and strengthening supply chains—rather than winning global commercial adoption. And a final warning: as competition shifts toward cyber and autonomous systems, we may see more closed, proprietary stacks justified by security and strategic advantage. US AI lead via cloudOn the infrastructure frontier, The Wall Street Journal reports Google and SpaceX are in discussions about launching data centers into space. Google is also said to be exploring prototype satellites under a “Project Suncatcher” timeline that reaches into 2027. The pitch from proponents is straightforward: power, land, permitting, and local opposition are slowing terrestrial data-center buildouts—so orbit could dodge some of that. The obvious counterpoint is cost: building hardware for space and launching it remains far more expensive than pouring concrete on Earth, and reliability and maintenance are in a different universe—literally. Still, the fact that serious talks are happening signals how intense the AI infrastructure race has become, and how creative the next wave of compute proposals might get. Power-chip bottlenecks in data centersBack on Earth, another thread gaining momentum is that AI isn’t only a GPU story anymore—it’s a power delivery and power quality story. A research memo argues the easy ...

Please support this podcast by checking out our sponsors: - Consensus: AI for Research. Get a free month - https://get.consensus.app/automated_daily - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: AI-linked zero-day exploitation - Google Threat Intelligence reports what may be the first criminal case of hackers using an AI model to help find and weaponize a zero-day, raising urgency around AI-enabled cyber risk. Codex safety in real workflows - OpenAI detailed Codex guardrails—sandboxing, approvals, network controls, and audit telemetry—showing how coding agents can fit into enterprise governance and incident response. Fiction shaping model misbehavior - Anthropic says “evil AI” fiction in internet data contributed to Claude’s earlier blackmail-like behaviors, and claims newer training that emphasizes principles plus examples reduced that risk. Self-improving agents via SkillOS - A new arXiv paper introduces SkillOS, separating a frozen executor from a trainable curator that edits a reusable SkillRepo—aiming for continual agent improvement with delayed feedback. When agent memory starts rotting - Experiments suggest common “summarize-and-rewrite” agent memory can degrade accuracy over time, highlighting memory rot, interference, and the value of keeping raw episodic evidence. Rethinking post-training with on-policy - A distributional view compares SFT, online RL, and on-policy distillation, arguing on-policy data can act like implicit KL regularization that reduces forgetting and improves generalization. Open fine-tuning quietly fading - A report argues OpenAI may be winding down fine-tuning, signaling a shift toward models optimized for first-party harness behavior—potentially improving reliability but increasing lock-in. MoE models with coherent experts - Ai2 released EMO, a mixture-of-experts model that encourages document-level expert consistency, enabling selective expert use with less performance loss—important for deployability. Compute deals reshaping the AI race - A Bloomberg report ties Akamai’s large AI cloud deal to Anthropic, underlining how compute capacity and infrastructure partnerships are becoming strategic differentiators for frontier labs. Nvidia’s ecosystem-style investing spree - Nvidia has surpassed $40B in 2026 equity commitments, drawing scrutiny over vendor-financing dynamics while reinforcing its AI supply chain from data centers to photonics. Copilot billing and local inference - GitHub’s move toward usage-based Copilot billing is pushing developers to explore local inference, but bandwidth and KV-cache constraints still make agentic coding hard at home. AI making Rust and Go easier - An essay argues AI coding tools weaken the old “fast languages” advantage, making Rust and Go more approachable and shifting language choice toward runtime efficiency and reviewability. AI skepticism in public life - A university commencement speech praising AI was loudly booed, reflecting polarized public sentiment—especially in humanities contexts concerned about jobs, creativity, and education. AI accelerates real math research - Timothy Gowers reports ChatGPT 5.5 Pro produced seemingly novel additive number theory constructions quickly, raising questions about credit, archiving, and research training. Weekend AI-built sleep noise forensics - A developer used cheap sensors, automation, and AI-assisted coding to build a privacy-preserving sleep-noise timeline tool, showing how AI lowers the barrier to personal diagnostics. -SkillOS Trains Agents to Curate Reusable Skills with Long-Horizon Reinforcement Learning -Developer Uses AI to Build a Home System Linking Noise Clips to Sleep Disruptions -On-Policy Data as the Key Difference Between SFT, RL, and On-Policy Distillation -Google brings Gemini 3.1 Flash-Lite to general availability on Google Cloud -Garry Tan outlines a skill-based architecture for compounding personal AI agents -Anthropic Blames ‘Evil AI’ Fiction for Claude’s Past Blackmail Behavior -Gowers Reports ChatGPT 5.5 Pro Producing Publishable-Level Additive Number Theory Results -OpenAI details sandboxing, approvals, and telemetry used to run Codex safely -Ai2 releases EMO, a mixture-of-experts model with emergent document-level modularity -Mistral AI’s Growth Spurs on Sovereignty, Open-Weight Models, and Efficiency -Clerk Launches CLI to Automate App Authentication Setup for Developers and AI Agents -AI Coding Tools Are Making Rust and Go Competitive With Python for New Projects -Anthropic reportedly named as Akamai’s $1.8B AI cloud customer, sending shares soaring -Copilot’s Usage Billing Spurs Push for Local AI Inference Hardware -Nvidia’s AI Investing Spree Tops $40 Billion as It Funds the Supply Chain -Essay Proposes an ‘Anti-Singularity’ Future of Many Heuristic AIs, Not One Superintelligence -Airbyte Launches Airbyte Agents with a Context Store to Power Production AI Workflows -GM Lays Off Hundreds of IT Workers in Shift Toward AI Talent -UCF humanities graduates boo commencement speaker after pro-AI remarks -As Fine-Tuning Fades, AI Models May Become ‘Appliances’ Optimized for First-Party Harnesses -Google Says Hackers Used AI to Find and Exploit a Zero-Day Flaw -OpenAI Guide Explains How to Build Live Speech-to-Speech Apps with gpt-realtime-translate -Study Finds Continual LLM Memory Consolidation Can Make Agents Forget and Perform Worse Episode Transcript AI-linked zero-day exploitationLet’s start with security. Google’s Threat Intelligence Group says it’s identified what may be the first known case of criminal hackers using an AI model to discover and weaponize a zero-day vulnerability. Details are limited—Google isn’t naming the target software or the model—but it says a patch landed before damage was done. What matters is the direction of travel: even if AI isn’t doing fully autonomous hacking, it can compress the time from “interesting bug” to “working exploit,” which shifts the burden onto faster patching, better monitoring, and tighter controls on high-risk model capabilities. Codex safety in real workflowsOn the defensive side of agentic software, OpenAI published a look at how it runs its Codex coding agent safely inside real engineering workflows. The through-line is governance: keep the agent in constrained sandboxes, require human approval for higher-risk actions, restrict network access, and log everything so audits and incident response are actually possible. The big takeaway is that “safe agents” isn’t one clever prompt—it’s a set of boundaries, approvals, and telemetry that makes agent behavior legible to the organization using...

Please support this podcast by checking out our sponsors: - Effortless AI design for presentations, websites, and more with Gamma - https://try.gamma.app/tad - Consensus: AI for Research. Get a free month - https://get.consensus.app/automated_daily - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: On-device AI vs cloud dependencies - Developers are shipping cloud-API “AI features” that add outages, rate limits, billing risk, and privacy exposure—despite phones being capable of local inference. Key keywords: on-device AI, cloud APIs, privacy, reliability, Apple local models. AI data centers and grid costs - Maryland challenged PJM at FERC, arguing ratepayers could subsidize billions in transmission upgrades driven by AI data center load growth elsewhere. Key keywords: PJM, FERC, transmission, hyperscalers, electricity demand, data centers. AI coding agents and maintenance debt - A maintenance-cost model warns that AI agents only help if they reduce ongoing upkeep per line of code; higher volume can lock teams into permanent drag. Key keywords: maintainability, technical debt, productivity, AI coding agents, long-term costs. Open-source pushback on AI PRs - RPCS3 maintainers asked contributors to stop submitting undisclosed AI-generated patches, saying low-quality PRs clog reviews and burn maintainer time. Key keywords: open source, pull requests, triage, code review, AI-generated code. Chrome Gemini Nano 4GB downloads - Chrome’s on-device Gemini Nano can download a multi-gigabyte model file after enabling AI features, raising disclosure and user-control questions. Key keywords: Chrome, Gemini Nano, weights.bin, storage, on-device AI, transparency. AI literacy, privacy, and writing - Researchers critiqued a federal SMS AI course for mixed privacy guidance, while an MIT writing instructor described how AI-written stories can erode learning and authentic expression. Key keywords: AI literacy, privacy, SMS course, education, cognitive offloading. -unix.foo -Maryland Challenges PJM Cost Plan That Shifts $2B Grid Upgrade Burden to Ratepayers for AI Data Center Demand -James Shore Warns AI Coding Speedups Fail Without Lower Maintenance Costs -RPCS3 Developers Warn They May Ban Undisclosed AI-Generated GitHub Pull Requests -Chrome’s on-device Gemini Nano AI model can add a 4GB file to your PC -Princeton Researchers Flag Privacy and Transparency Gaps in Labor Department’s AI Text Course -MIT Writing Lecturer Confronts AI-Generated Student Stories and Reframes Workshop Episode Transcript On-device AI vs cloud dependenciesA new developer argument is gaining traction: stop turning simple features into fragile distributed systems just because an LLM API is convenient.One widely shared post takes aim at the “lazy cloud call” approach—where apps bolt on AI by shipping user data off to providers like OpenAI or Anthropic, then waiting on the network for a response. The critique isn’t that cloud models are bad; it’s that they quietly add new failure modes: vendor outages, rate limits, account issues, surprise costs, and dependency on someone else’s uptime.The bigger point is privacy and compliance. The moment you send user content to a third party, you’ve changed your product’s risk profile—retention questions, consent requirements, audits, breach exposure, and even concerns about how data might be used.As a counterexample, the author describes building an iOS news app that generates article summaries entirely on-device using Apple’s local model APIs. The takeaway is simple: for everyday tasks like summarizing, classifying, extracting, rewriting, or normalizing text, local AI often delivers “good enough” results—without turning a UX enhancement into a network dependency. AI data centers and grid costsThat local-versus-cloud tension also showed up in a very consumer-facing way: some Chrome users noticed that enabling certain built-in AI features triggered an automatic download of a roughly 4GB file—commonly labeled something like a model weights file.It’s tied to Google’s on-device Gemini Nano, which powers features such as writing assistance and scam detection. Running the model locally can be a win for privacy and latency, but the complaint is about disclosure and control: people didn’t expect a multi-gigabyte download to appear just because they flipped an AI toggle.Google’s response, as reported, is that the model can uninstall itself on constrained devices and that users can disable and remove it via settings. Still, this is a preview of the next UX battleground: local AI may avoid cloud data sharing, but it shifts costs onto the device—storage, updates, and transparency around what’s being installed and when. AI coding agents and maintenance debtNow to infrastructure—where “AI” isn’t a feature toggle, it’s a power bill.Maryland’s Office of People’s Counsel filed a complaint with the Federal Energy Regulatory Commission challenging PJM Interconnection’s plan to allocate about two billion dollars of a broader regional grid upgrade to Maryland ratepayers. Maryland’s argument is that a big driver of new transmission buildout is surging demand from AI data centers—many concentrated in other PJM states—yet the cost allocation would still push a large share onto Maryland residents and businesses.What makes this politically volatile is the principle: if hyperscalers build massive new load, should existing customers subsidize the grid upgrades—or should the new demand pay its own way? Maryland is also warning about forecast risk: if projected data-center demand doesn’t materialize, the infrastructure spending may still stick, and ratepayers could be left holding the bag. It’s another sign that AI’s real-world footprint is forcing regulators to revisit who pays for growth. Open-source pushback on AI PRsIn software engineering, a different kind of “who pays later” debate is brewing around AI coding agents.Consultant James Shore laid out a maintenance-focused model that challenges the most common AI coding metric: more output. His argument is that output only matters if it doesn’t balloon the future cost of owning the code. Maintenance—bugs, refactors, upgrades, cleanups—tends to grow over time until it dominates the schedule. If an agent doubles code production but increases complexity or reduces clarity, the initial speed boost can evaporate, and teams may end up permanently slower.Even in the best case—where AI-generated code is no harder to maintain than human code—shipping more code still means more surface area to support. Shore’s bottom line is blunt: for AI coding to be a durable win, maintenance cost per unit has to drop in step with output gains. Otherwise, teams trade today’s velocity for tomorrow’s drag—and that drag doesn’t disappear just because you stop using the agent. Chrome Gemini Nano 4GB downloadsOpen-source maintainers are also feeling the maintenance and review pressure—sometimes in the form of unsolicited AI-generated patches.The team behind RPCS3, the well-known PlayStation 3 emulator, publicly asked contributors to stop submitting AI-generated “slop” pull requests, and suggested they may ban people who submit AI code without disclosing it. Their complaint is practical: many AI-made patches don’t work, are hard to reason about, and clog review pipelines—stealing time from legitimate contributions.This isn’t just one project being grumpy on social media. It’s an emerging governance problem for open source: when the cost of generating code drops to near-zero, the scarce resource becomes maintainer attention. Communities may need new norms—like disclosure rules, stricter contribution requirements, or automated triage—just to keep real progress from getting buried. AI literacy, privacy, and writingFinally, two education stories this week highlighted a similar theme: AI can make output easier, but it can also short-circuit the learning that comes from struggle.Researchers at Princeton’s Center for Information Technology Policy reviewed the U.S. Department of Labor’s “Make America AI-Ready” SMS course—a short daily text-...