Anjney Midha

Anjney Midha argues that money and compute do not guarantee shipping; aligned execution and culture are the constraints. He says Google treated roughly 95% node utilization as outage territory, while best-in-class MFU is 60–70%; poorly coordinated labs let small planning errors compound across organizational layers. “AI scaling should be putting a premium on the value of common sense.”

Amp’s proposed answer is a neutral compute grid that pools fragmented supply across clouds and silicon. The goal is to make “megaflops flow like megawatts,” guaranteeing participants base capacity while dynamically allocating spikes. Amp has started securing demand at a stated 1.3-gigawatt scale—roughly $40 billion of cloud spend—but Midha estimates its teams could need 6 gigawatts of spike capacity over four years.

Data centers need an explicit bargain with their host communities or permitting risk becomes infrastructure risk. Swyx relays an estimate that up to 20% of U.S. data centers this year may lack sufficient community support, though he cautions it could be overstated. Scott Nolan’s proposed “new AI deal” would raise compute from $4 to $4.50 an hour and return the incremental $0.50 directly to locals through cash; Swyx also suggests cheaper electricity. As a compute customer, Swyx says he would gladly pay.

Alternative AI chips can expand supply without fragmenting every layer of the stack. Matrox chose NVIDIA’s reference architecture and rack footprint, then concentrated innovation on the logic die and systems co-design: “You just can’t fight on every front.” The deeper constraint is trust—chips take roughly two years to tape out, so designers need early visibility into changing model architectures.

Research hoarding inside vertically integrated labs creates an opening for independent capital and infrastructure. Midha says DeepMind’s six-month internal embargo can become permanent when work appears commercially useful, producing an adverse-selection problem in what gets published. Amp’s Foundry arm therefore backs frontier teams—including a stated few hundred million dollars invested in Anthropic earlier this year—while Amp also donates excess compute to nonprofits and university labs.

End-of-life prediction is Midha’s clearest example of outputmaxxing with public value. He says Stanford’s longitudinal dataset covered at least 12 million patients, while more than 30% of Medicare and Medicaid spending at the time went toward end-of-life care; even regression and simple neural nets appeared technically useful. Regulation, not modeling, remained the block because liability could not shift from physicians to an AI system: “I haven’t been able to get this out of my mind a single day for the last 14 years.”

Culture is the fragile compounding advantage behind Anthropic’s coding breakout. Midha rejects the “lucky dice roll” explanation: Anthropic spent four years becoming prepared under scarcity, made coding its P0 because computer-use capability could advance AGI, and endured “21 noes” that forced clarity. “Culture is not a set of beliefs. It’s a set of actions”—a garden requiring daily reinforcement, not a permanent moat.

🔗 Original source & video: Why AI Labs With Unlimited GPUs Still Fail — Anjney Midha, AMP

AI funding has already outgrown venture scale: Peter Diamandis put U.S. deployment at $1 billion a day, potentially reaching $3 billion a day by 2030—and said he expects it to exceed that—while Dave Blundin compared it with roughly $200 billion of annual U.S. venture investment. Anjney Midha’s answer on how much a16z capital is flowing toward AI was “basically, all of it, and it’s still not enough.” Strategic investors, public markets, credit and sovereign capital all need to join the funding effort.

The compute preference stack runs from raw cash to GPUs to high-quality foundation-model tokens, while electricity is becoming the hard infrastructure constraint. Reasoning models generate roughly 10 times more tokens than earlier generative-AI models, producing a daily Jevons paradox in which efficiency unlocks still more demand. NVIDIA’s Blackwell NVL72 may reach data centers before they can be cabled, permitted and powered: “We just don’t have enough electricity to power the chips.”

Public markets are becoming part of the funding engine, with Bonnie Chan saying Hong Kong ranked first in that year’s global IPO league table and had completed about 80 deals with 300 more in the pipeline; about half of the combined group involved AI. HKEX can draw demand from technology-enabled “pro-retail investors,” but Chan warned that enthusiasm will eventually give way to harder valuation questions. Funding AI will require matching opportunities with private, public, credit and equity capital.

Vertical applications offer the clearest early-stage economics because their use cases are abundant and their capital requirements are lower than foundation models or data centers. Blundin said qualifying MIT and Harvard teams had so far achieved a near-100% success rate; typical entry valuations of $20 million-$30 million can be followed by $100 million-$300 million in first funding, and a company that is going to become a unicorn can get there within two years. Mercor’s progression—from $30 million at founding to $300 million, $2 billion and $10 billion—was his emblematic case.

The panel’s most serious risk was political: frontier-AI wealth is compounding privately while power costs, job disruption and infrastructure trade-offs reach the public. Midha cited Anthropic rising from a few hundred million dollars in valuation to $183 billion in 48 months, then asked, “Where’s my piece of the future?” He warned that India could face an “ugly” transition if Claude and GPT-5 tokenize vast portions of its IT-services flow. Peter distinguished that future shock from current layoffs driven by 2010-2020 over-hiring, while Midha added the COVID-era print-money period.

The proposed access mechanism is institutional stewardship—not pushing retail investors into opaque, already-high valuations. Midha wants sovereign, pension and state funds on frontier-AI cap tables; Anthropic’s seed round received 21 rejections from 22 introductions before being pieced together from angels and high-net-worth individuals. Blundin separately warned that capital-intensive robotics and fusion, along with speculative bets such as quantum computing, could sour confidence in genuine AI value creation, as peripheral bets did around the internet crash.

🔗 Original source & video: AI Investor Panel: Where Smart Money Is Actually Going in AI | EP 219

Periodic Labs’ $300 million seed, led by Andreessen Horowitz, backs a capital-intensive thesis: an AI scientist must learn by acting on physical reality, not merely by absorbing the internet. Liam Fedus calls experiment a “physically grounded reward function,” with simulations and language models as tools but nature as the final error-corrector: “Nature is our RL environment.” The investable proposition is a lab-built loop of hypotheses, automated experiments, positive and negative outcomes, and improved models.

Scaling laws may continue to hold while still failing to deliver useful scientific discovery on the required timescale. Fedus’s challenge is “what is this y-axis?”: scaling against internet or coding distributions does not manufacture missing physics knowledge, and “that model is not going to then cure cancer.” Ekin Dogus Cubuk adds that out-of-domain performance can improve as a power law yet have such a shallow slope that reaching the target might take centuries.

The existing scientific corpus is not just too small; it is noisy, selectively published, and missing the iterations that teach scientific judgment. Reported physical properties can span orders of magnitude, formation-enthalpy errors can defeat prediction, and superconductivity datasets have a high noise floor. Because negative results are rarely published, a model trained on literature can at best reproduce a distorted distribution rather than learn why an experiment failed.

High-temperature superconductivity gives Periodic a measurable north star and forces it to build the entire autonomous-science stack. The stated ambient-pressure benchmark is about 135 Kelvin; exceeding it would provide an unambiguous score, while a hypothetical 200 Kelvin superconductor would, Cubuk argues, update humanity’s view of the universe even before commercialization. Getting there requires autonomous synthesis, characterization, simulation, and experiment selection—capabilities that can be tested for transfer into magnetism and other physical domains.

The near-term business is an intelligence layer for advanced manufacturing, not a wait for a miraculous superconductor. Fedus targets copilots for researchers and engineers in semiconductors, space, defense, and other industries with “massive R&D budgets,” reducing iteration time across literature review, simulation, design, and experiment. Deployment follows a land-and-expand motion: solve one critical, well-scoped problem with clear evaluations rather than promise to transform an entire fabrication line on day one.

Periodic’s model strategy goes beyond retrieval by encoding private scientific and industrial knowledge through mid-training and high-compute reinforcement learning. Mid-training means continuing pre-training on knowledge absent from the base model—from crystal structures and simulation outputs to descriptions of how materials were made—then connecting those distributions so one dataset improves performance on another. That deeper encoding creates an enterprise challenge too: knowledge may need to be bucketed into separate systems when some data is accessible only to senior leadership.

The organizational moat is a roughly 30-person, cross-disciplinary team built around curiosity, translation, and urgency rather than credentials. Weekly teaching sessions let ML researchers explain RL and data cleaning while physicists and chemists teach quantum mechanics and scientific history; the cultural rule is “no stupid questions.” Advanced degrees are explicitly unnecessary because even the best specialist knows far less than the combined physics, chemistry, synthesis, and characterization the mission demands—and the founders want progress “ASAP,” not in 10 years.

🔗 Original source & video: Periodic Labs: Training AI Scientists, with Liam Fedus & Ekin Dogus Cubuk (from a16z)

Periodic Labs’ central wager is that nature should become the next reward function for AI. Early ChatGPT learned helpfulness from human preferences and later gained mathematical and coding correctness through verifiable graders; advancing physics requires the same optimization pressure against real experiments. Cubuk’s formulation is the thesis in one line: “Nature is our RL environment.”

More compute will improve models, but it cannot manufacture missing scientific knowledge or efficient out-of-domain learning. Fedus accepts that scaling laws continue to hold, then asks, “What is this y-axis?” A coding model can recursively improve at passing unit tests, but “that model is not going to then cure cancer”; Cubuk adds that an out-of-domain power law may have such a shallow slope that progress would take “centuries.”

Periodic is building the data engine that the scientific literature cannot provide. Published measurements can span orders of magnitude, negative results are rarely reported, and synthesis or superconductivity datasets may have noise floors too high to train predictive models. Because “these systems aren’t magic,” experiment must collapse uncertainty and continually move the training distribution toward the target.

High-temperature superconductivity is both a falsifiable benchmark and a forcing function for the full autonomous-science stack. The ambient-pressure mark cited is roughly 135 Kelvin; the founders say exceeding it would likely require autonomous synthesis, characterization, simulation, and experimental iteration. A hypothetical 200 Kelvin superconductor would matter even before commercialization because observing quantum effects at that temperature would be “such an update to people’s view of how they see the universe.”

The commercial wedge is an intelligence layer for engineers and researchers in space, defense, semiconductors, and advanced manufacturing. Periodic wants systems that automate simulations, connect design pipelines, and reduce physical R&D iteration time across “massive R&D budgets.” Fedus explicitly links mission and economics: “Technology and capital are intertwined,” so the lab plans a scoped “land and expand” motion rather than attempting to transform a production line on day one.

Mid-training—not retrieval alone—is how Periodic expects to turn general models into physics and chemistry experts. It plans to continue pre-training on crystal structures, synthesis recipes, simulations, experiments, and customer knowledge, then use high-compute reinforcement learning and specialized tools. The organizational design mirrors this composition: a roughly 30-person “N of one” team spanning LLMs, experiments, simulations, automation, and theory, reinforced by academic advisers and grants.

🔗 Original source & video: Building an AI Physicist: ChatGPT Co-Creator’s Next Venture

GPT-5’s strategic purpose is to make reasoning the default rather than force users to choose between instant GPT models and the slower o-series. OpenAI is researching how much thinking each prompt deserves, aiming to remove that product friction while delivering more agentic behavior “by default.” For investors, the competition is shifting toward adaptive thinking budgets, latency, reliability, and usable autonomy.

OpenAI considers many familiar evals effectively saturated and is moving toward benchmarks based on genuine discovery and economic relevance. Reinforcement learning can create narrow domain experts, so improving from 96% to 98% may say less about generalization than it once did. AtCoder and IMO remain credible markers because leading researchers passed through them, but the next milestone is “actual movement on things that are economically relevant.”

The central research roadmap is an “automated researcher” capable of discovering new ideas in machine learning and other sciences. Jakub Pachocki estimates that reaching near-mastery of high-school competitions would correspond to roughly “one to five hours of reasoning”; progress now requires longer planning, persistent memory, recovery from failed approaches, and autonomous operation measured over expanding time horizons.

Reinforcement learning keeps producing gains because language-model pretraining supplied the rich environment that earlier RL systems lacked. Natural-language modeling gave models a nuanced understanding of human language, after which researchers could explore many objectives and domains. Mark Chen expects reward design to become simpler, while Pachocki says learning should move toward something more humanlike and warns enterprises “not to assume that what is now will be forever.”

GPT-5-Codex shows that deployment quality depends on allocating intelligence and time correctly, not merely maximizing it. The previous generation spent too little time on the hardest tasks and too much on easy ones; the new work targets lower latency for simple jobs and deeper reasoning for difficult, messy coding environments. Jakub Pachocki, who said he had mostly used Vim, said a 30-file refactor can now be completed “pretty much perfectly in 15 minutes,” although the tools remain in an “uncanny valley” short of a coworker.

“Vibe researching” is a hoped-for future, but the guests argue that taste, persistence, and honest failure analysis remain load-bearing. Research means attempting something “that will most likely fail,” maintaining conviction without going out of one’s way “to prove that it works,” and recognizing both software bugs and flawed conceptual frames. The human research contribution still includes choosing important, hard problems and learning when to persist or pivot.

OpenAI’s organizational approach combines protected fundamental research, deliberate prioritization, and still-scarce compute. The lab resists chasing every competing release, maintains distinct mandates for algorithmic advances and product-oriented research, and Jakub’s interrupted answer to an extra-10%-resources question was “compute.” The risk of diffuse investment is ending up “second place at everything,” while longer-term constraints broaden from compute to energy, robotics, and the physical world.

🔗 Original source & video: From Vibe Coding to Vibe Researching: OpenAI’s Mark Chen and Jakub Pachocki

Genie 3’s significance extends beyond better video: it is a new kind of model for interactive worlds generated in real time from a few words. Its one-minute spatial memory, frame-to-frame consistency and immediate controls turn passive clips into navigable environments—“there is something magical about the real-time aspect.” It remains a research preview, with broader access intended but no concrete timeline.

The leap came from combining capabilities previously split across Genie 2, Veo 2 and GameNGen. The team pursued the most ambitious intersection—higher resolution, real-time generation and “minute-plus memory” in one model—despite those objectives conflicting. The result arrived after roughly seven months and resonated more strongly than its creators expected.

Spatial persistence was explicitly designed, yet its quality still surprised the researchers who built it. Genie 3 generates frame by frame without an explicit NeRF, Gaussian-splatting or other fixed 3D representation; the team believes that choice is key to generalization. The present design retains this memory for one minute, although Shlomi says there is “no fundamental limitation.”

Scale and training breadth are yielding increasingly credible physical behavior, though the researchers stop short of calling it LLM-style reasoning. In examples, characters typically swim when entering water, skiing speeds up downhill and slows or stops uphill, and an approached door may open; non-experts can mistake some storms, lighting and water for reality. The model finds it harder to obey unlikely prompts while preserving world coherence—“low-probability areas” such as wearing flip-flops in the rain.

Genie 3 and Veo 3 remain separate because interactivity and cinematic generation impose different technical priorities. Genie offers navigation and actions but generally lacks audio; Veo 3 targets a higher visual-quality threshold, while agent training values rapid, egocentric interaction over cinema-grade output. Shlomi frames modality, generation speed and controllability as orthogonal dimensions rather than a single inevitable convergence path.

A promising path is synthetic experience for agents and robotics. Anjney highlighted a possible composition with an agent she thought was called SIMA; Jack says Genie 3 is an environment rather than an agent, so other agents can learn through simulated experience. This could combine real-world data’s realism with simulation’s scale and safety—the “best of both.” Simulation still does not solve actuation, movement decisions or the broader physical-response loop.

The team simultaneously describes Genie 3 as years ahead of prior expectations and far from an accurate world simulator. Jack says minute-long, photorealistic, remembered worlds looked like a five-year goal only two or three years ago; Shlomi cautions that genuinely placing a person or agent into a faithful world requires much more work. Calendar forecasts remain deliberately hedged because “we live in an accelerated timeline.”

🔗 Original source & video: Google DeepMind Lead Researchers on Genie 3 & the Future of World-Building

Washington’s AI posture has flipped from constraining AI amid existential-risk warnings to building a platform the United States intends to lead. The action plan opens with “a new frontier of scientific discovery,” backs open source, and proposes an evaluations ecosystem to measure risk before making grand proclamations. The guests see this as a cultural shift from “PauseAI” toward empirically measuring risks and opportunity costs.

SB 1047 became the cautionary example of how theoretical harms can create immediate commercial chilling effects. The proposal could have exposed open-weight developers to downstream liability if someone later fine-tuned their work and caused a mass-casualty event; Erik Torenberg recalls one version encompassing three deaths or an overwhelmed medical system. For a developer who “can’t even afford” litigation, merely moving the question into court can suppress experimentation.

DeepSeek punctured the premise that restricting American open source could preserve a multi-year lead over China. DeepSeekMath-V2 had already signaled proximity to the frontier before R1 surprised Washington, while distillation meant the marginal advantage from withholding weights was limited. Erik Torenberg’s blunt challenge: “Have you actually looked at the author list of any paper in AI?”

Open weights now have a concrete business case extending well beyond open-source philosophy. Closed models can pioneer frontier capabilities while open models serve governments, regulated industries, and Fortune 50 customers demanding on-prem deployment, control, security, and support—the emerging “sovereign AI market.” Because weights do not include the underlying data and training pipeline, companies can distribute smaller models while retaining larger paid models and core IP.

The investable market may bifurcate rather than converge on one winning licensing model. Frontier APIs and controlled deployments address different customers, infrastructure, support requirements, and revenue models; the guests expect winners in both. Waiting for the structure to settle is itself risky when founders in their twenties can build businesses with revenue run rates in the “tens to hundreds of millions of dollars” within a few years.

The action plan’s direction is stronger than its implementation detail, with academia the conspicuous omission. Its call to “build an AI evaluations ecosystem” replaces proclamation with measurement and quickly became a reference point for other governments. Yet Anjney Midha argues that pursuing a major technology initiative without universities leaves the country fighting “with a hand tied behind our back.”

The guests reject the idea that incomplete interpretability justifies waiting indefinitely at the frontier. Models may be “grown, not coded,” but society routinely extracts value from complex systems it cannot explain atomistically; alignment can improve usefulness without requiring a universal ideological mandate. Their opportunity-cost framing is categorical: “The p(doom) without AI is actually quite a bit greater than the p(doom) with AI,” especially if delay slows disease and scientific discovery.

🔗 Original source & video: The Current Reality of American AI Policy: From ‘Pause AI’ to ‘Build’

Sovereign AI is not a settled architecture but a demand for control over technical, legal, and cultural dependence. For enterprises, that means knowing where workloads sit and which governments can compel access; for countries, it means ensuring models used across critical industries and daily life do not encode an adversary’s values. Anjney Midha’s broad definition is “control over your own destiny as much as possible as AI plays out.”

Enterprise AI is rebundling technology, implementation, and insurance—and jurisdiction may determine who wins the bundle. Hyperscaler bundles can absorb security and compliance risk through indemnification, while Midha’s CMA CGM example includes cybersecurity and copyright exposure; he argues European buyers may prefer Mistral when the US CLOUD Act makes an American provider unacceptable. The purchasing question is therefore not merely who has the best model, but “are they buying technology or are they buying insurance?”

Most countries cannot compete for frontier talent today, but Midha rejects the claim that they can never catch up. His build-buy-partner framework favors immediate access through joint ventures, medium-term investment in “forward-deployed AI solutions engineers,” and a 10- to 20-year path toward local pre- and post-training capability. With researchers bid against Meta and other frontier labs, sovereign balance sheets may be needed to seed that ecosystem.

Every country should tokenize its culture, then separate commodity pre-training from the locally defensible last mile. Nathan Labenz argues governments should hand curated linguistic and cultural corpora to every leading developer; Midha agrees they should partner for pre-training, preferably through open weights, while retaining local productization, continual post-training, distribution, and non-verifiable reward design. His test is personal: ChatGPT in Hindi sounds “like an American tourist visiting India,” not a local.

Open models currently trail the closed frontier by months rather than years, making sovereignty through post-training plausible—but contingent. Midha cites 26 days from o1 to DeepSeek R1, characterizes China’s reasoning fast-follow window as within 60 days, and puts the broader open-versus-closed lag at roughly six months or less. He expects China to keep open-sourcing frontier models as soft power, though Labenz stresses that this rests on strategic choices by a few state-backed or corporate actors, not a conventional open-source community.

AI factories are rewriting cloud economics and creating a strategic contest between custom silicon and Nvidia-backed “open scalers.” Midha says GPUs have risen from under 10% to roughly 60-70% of a data center’s bill of materials; Amazon’s lack of a GPT-4 alternative helped drive an $8 billion Anthropic investment as contracts worth more than $100 million moved toward Azure. Gemini–TPU and Anthropic–Trainium integration threaten Nvidia, while Nvidia answers through CoreWeave, Nebius, Mistral Compute, and regional sovereign clouds.

The Middle East offers the US capital, faster construction, cheaper energy per FLOP, and geopolitical alignment—but infrastructure commitments compound for decades. The discussion says relevant deals require one-for-one infrastructure investment in the US, and Midha estimates a liquid-cooled Blackwell node can deliver roughly 18-20% lower energy cost per FLOP in the UAE or Saudi Arabia, absent US subsidies. His strategic warning is that “infrastructure is destiny”: losing the first three to five years to Huawei could determine the next 30, even if China cannot supply equivalent systems today.

🔗 Original source & video: Sovereign AI: Geopolitical Strategy & Industrial Policy for Countries 3-193, with Anjney Midha, a16z

Sovereign AI is breaking the cloud-era assumption that global workloads will concentrate in the US and China. The kingdom announced HUMAIN, a local hyperscaler or AI platform intended to run most AI workloads domestically, within an announced cluster buildout somewhere around $100 billion-$250 billion; roughly 500 megawatts appears to be the “atomic unit.” The goal is “infrastructure independence,” including autonomy over models and deployment.

“AI factories” are technically distinct from conventional data centers. GPUs are the major active-component difference, while high-density clusters require rack-level liquid cooling, an energy supply close to a power plant, and early energy commitments. Enterprises may also bypass elaborate cloud stacks for Kubernetes plus selected Snowflake- or database-type services.

Models have become cultural and information infrastructure, making foreign dependence a national vulnerability. Training data embeds values, while post-training steers what models answer or refuse; meanwhile, foundation models already touch defense, healthcare, finance, and the daily decisions of ChatGPT’s roughly 500 million monthly active users. As models replace search and grade schoolwork, whoever controls them could shape accepted history and truth.

AI data centers resemble industrial-era oil reserves—with the crucial difference that countries can construct them. Capital and political will can create the compute base upon which domestic industries, development, and exports are built.

The US faces a choice between helping allies build sovereign capacity and leaving the field to Chinese models. Midha’s preferred analogy is a “Marshall Plan for AI”: the original reconstruction looked like a capital export but produced a 70-year US-Europe trade corridor and kept China out of that equation. At the model layer, he reduces the diplomatic choice to “DeepSeek or Llama?”

Appenzeller rejects comprehensive government control while preserving a targeted government role. Government can fund fundamental research and set sound regulation, but competitive companies must supply the detailed innovation; even the Manhattan Project leaked, making total control “a pipe dream.” DeepSeek’s MIT-licensed release—26 days after OpenAI’s frontier release—supports his strategy of building and exporting the best technology, ideally from the US and its allies, while Midha calls the resulting approach “foundation model diplomacy.”

🔗 Original source & video: Sovereign AI: Why Nations Are Building Their Own Models

AI sovereignty has moved from an adoption debate to a build-or-buy decision. Anjney Midha argues that billions already use AI, leaving governments little practical choice over whether to embrace it. The host frames nation-state spending over the next 24 months as potentially “the single largest purchasing decision.”

Smaller countries do not need the entire AI stack, but they must choose trusted partners. Midha divides the world into frontier “hypercenters” and “compute deserts”; joint ventures can provide infrastructure independence without full ownership. Because training data embeds cultural norms, alignment also means deciding “whose values align more with yours.”

Compute, energy, data, and regulation are the four ingredients Midha says really matter. Resource-rich countries can trade their comparative advantage for missing capabilities—for example, Middle Eastern energy could attract model labs and technical talent. Jointly trained models may provide “joint independence from a value system that you don’t subscribe to.”

True sovereignty means controlling critical dependencies, not replicating every layer. ASML’s roughly $200 million lithography machines illustrate the constraint: rebuilding that capability in the US could take “10-plus years,” while training a local frontier model might take months or quarters—if a country can secure one of the few capable research teams.

Midha sees US data, energy, and inference-liability policy as key risks, while saying private compute markets are doing pretty well. He estimates that more than 700 pieces of state-level AI legislation appeared in 2024, creating a changing patchwork where companies cannot tell “what they should even comply with.” He also argues that constrained nuclear development and proposals making model developers liable for downstream misuse could drive startups abroad and entrench Big Tech.

The strongest signals are sovereign GPU orders and technically credible founders. Governments are ordering NVIDIA GPUs 12–36 months before delivery because the data center is becoming “the new atomic unit of sovereignty.” Midha is watching deeply technical founders with frontier-model experience—such as Arthur Mensch, who started Mistral AI, and Guillaume Lample, who led the initial Llama family at Meta—willing to build infrastructure for governments, a role whose impact could be “quite generational.”

🔗 Original source & video: AI Is Becoming a Regional Race