DEMO 01/03PREVIEW
Podcasts are full of confident predictions that evaporate the moment the episode ends. The graph keeps them — 393 dated, falsifiable calls from 87 people, each one attributed, timestamped, and quoted verbatim. When the deadline passes, the receipt is right here.
→ the latest call with a deadline, printed
The Anthropic IPO will create roughly 100 new EA billionaires whose political influence will be like George Soros times 100 or 1000.
Prediction volume by person, counted across every indexed show. Some people hedge; these 8 commit.
245 of 393 carry an explicit deadline — checkable calls. The rest are open-ended, and the ledger remembers those too.
Filter by show or keyword — every row links back to its source, to the second.
The oil concession will improve Venezuela's economy.
States and cities that embrace AI, data centers, tax cuts and teacher bonuses will pull ahead while far-left cities hollow out, reordering the US economic and voting balance.
The Anthropic IPO will create roughly 100 new EA billionaires whose political influence will be like George Soros times 100 or 1000.
Anti-AI and anti-data-center positions taken now will look incredibly dumb as the truth about the backlash is exposed.
Shapiro and Abbott will switch back to being pro-data-center after the midterms.
Most conservatives and libertarians will come out in favor of open AI as the open-vs-closed debate evolves.
Nvidia, as the leading American open-source provider, will catch up to and blow past the Chinese models.
The Anthropic and OpenAI IPOs will fund more than half a trillion dollars of charitable giving and donor-advised funds.
As software infrastructure becomes dynamic and agentic, cyber defense will become as strong as or stronger than offensive agent systems.
The next ten years will bring continual security incidents while legacy code is replaced, after which agent-written code exceeds human-written code and AI offense and defense reach a stalemate.
San Francisco's ultra-luxury housing market could reach $5,000 per square foot.
The market may reach 1999-style peak euphoria by the end of 2026.
The AI euphoria will continue roughly three more years before the cycle turns.
Anthropic and OpenAI will eventually ship their own Grokbot-style bot products.
GPT-6-class capabilities will be matched or nearly matched by a handful of closed-source and open-source alternatives.
Personalized cancer immunotherapy proliferates overseas at much lower cost over the next couple of years via right-to-try and cheap clinic processes.
In five years the large tech companies will each have their own cloud, models, silicon, and data centers — soup to nuts.
AI capex continues well into the future rather than ending soon; Nvidia's guide is the tell.
Between 2030 and 2032 Social Security cannot pay and multiple states go bankrupt, forcing Social Security and state bailouts/restructurings.
If the long-term Treasury yield prints 6%, that is the beginning of a years-long fiscal death spiral unless Congress gets the budget under control.
By the end of this decade, a single frontier lab holds more effective AI labor than there are people on earth.
In the 2030s the rate of interest reaches tens of percent as an automated economy potentially doubles every year.
Heavily-indebted, low-tax-revenue countries such as Pakistan and Nigeria are hit hardest by the new interest rate regime.
A second Volcker shock repeats the 1980s pattern, defaulting a wave of heavily-indebted developing countries.
Meta's borrowing cost rises to around 8%, up from the 5-6% it raised at recently, dragging the whole economy's cost of credit up with it.
AI CapEx from 2024 to 2029 totals about $11 trillion, roughly $6T cash-funded and $5T debt-funded.
Incremental global AI CapEx approaches $10 trillion a year by the end of the decade.
China can add 50 gigawatts of incremental compute in 2029, though quality-weighted that is worth closer to 20 gigawatts of American chips.
Chinese domestic fabs (SMIC, CXMT) add 5-10 gigawatts of domestically produced compute in 2028 alone.
China will have roughly 30 gigawatts of AI compute or less in 2028.
Global incremental AI compute additions reach 90-100 gigawatts in 2029.
By the end of 2027 a leading lab's blended revenue per megawatt reaches $70-80M or higher.
Most compute will still transact at under $20 billion per gigawatt at the end of 2027, despite the headline $40B SpaceX deal.
By the end of 2028, OpenAI and Anthropic together control most of the world's usable flops.
By the end of 2027, half of the world's incremental new compute goes to just OpenAI and Anthropic.
AI will not replace enterprise sales in the near term because the work depends on human rapport and improvisation.
Roughly 35-40% chance of something we would recognize as AI takeover by 2040.
On his default modal timeline, the misalignment situation becomes really crazy and concerning about three years from now.
AIs sufficiently good at hardware R&D, fabs and robotics produce an industrial explosion, building out far more compute and physical capacity.
Getting five years of AI progress in one requires roughly eight years' worth of algorithmic progress, to also cover the missing compute scale-up.
Automating a video editor's job lands around the same time as full automation of AI R&D, well before all human jobs.
AI beats all humans on the job — the general-competence milestone — with a median around 2033.
Full automation of AI R&D arrives around 2030-2031.
Automating AI R&D yields a median of four to five years of AI progress in a single year.
The Taiwan issue will most likely be resolved through economic and political pressure leading to reunification with China, rather than through a full-scale US-China war.
The most likely beneficial outcome of the current world-order transition is a more regional arrangement, with the US and China each dominant in their own region rather than one global superpower.
In roughly 20 years, the approximately 9 million human drivers doing rides and deliveries today could be replaced by around 20 million autonomous vehicles.
The realisation of scarcity will arrive slowly and then all at once, ending in a mad scramble as inventories collapse.
Gold will continue to outperform the stock market over the long haul because money printing will accelerate.
Europe will be crushed this coming winter for lack of natural gas to heat cities like London.
China will bail Japan out, neutralising the Japan problem and serving Beijing's interest in a dependent, non-nuclear Japan.
Japan is his dark-horse candidate for triggering a broader financial crisis, as a weakening yen and lost control of bond yields force market closures and treasury sales.
Even if the Strait reopened immediately, flows would not return to 21 million barrels a day for about a year, with a minimum of three to four months just to untangle shipping.
China will have to intervene in some meaningful way if the Gulf disruption continues, because it cannot accept the region being shut down for a year.
Inventories will be run down to scary minimums and then government rationing will be attempted.
Oil could easily reach $300 to $400 a barrel, because it is severely underpriced relative to the work it performs.
Gasoline will reach $6 a gallon before Labor Day or later this year.
Real hope for the Iranian people is five to ten years away, and continued escalation is pushing that horizon closer to ten years than five.
If Iran consolidates its Hormuz control and sphere of influence, it is very likely to acquire nuclear weapons within a year or two.
The Strait of Hormuz will almost surely not be physically reopened before the midterm elections.
In the next few months the US will cross the threshold into ground operations on an Iranian island or coastal territory, as the world economic crisis and US gasoline prices keep growing.
If a mass-casualty attack against Americans occurs before the midterms it will raise Trump and the Republicans' odds of winning; without such an attack, a worsening economic crisis makes a Republican loss likely.
The biggest near-term risk is a mass-casualty attack against American troops (a ballistic-missile barrage on naval ships in the Gulf of Oman) or American civilians.
There is roughly a 70% chance the US carries out a limited ground operation in Iran, likely within the next month and possibly by the end of this week.
As utilities and gas generators begin hedging and contracting physical gas supply for 2028, a 'knife fight' to secure natural gas will emerge that hasn't been seen before, finally moving the currently flat forward curve.
By 2029, natural gas could rise to 20-30% of hyperscalers' cost of compute, up from about 10% today, as gas prices double or triple structurally.
Orders for new large-scale natural gas power generation could slow very meaningfully as 2026 progresses, since it may not make sense to build new gas plants past 2029-2030 without secured, deliverable gas supply.
New large-scale nuclear reactors (AP1000 units) will need to come online by 2033-2034 — the earliest realistic timeline — to address the structural gas and electricity deficit as it worsens into the early 2030s.
Residential solar adoption will grow exponentially from here, even without federal tax incentives, as households seek protection from rising peak electricity prices.
The natural gas price risk becomes unbounded and convex as the structural deficit takes hold in the back half of the decade, felt most acutely in electricity prices in 2028-2030.
Starting in the middle of 2028, the US will draw down its working natural gas storage in a historic way; by 2029 storage will drop below all known historical levels, and by 2030 it will approach critically low levels.
In an unmitigated (P30/P0) scenario, AI-driven incremental natural gas demand could more than double to 12-15 BCF/day by the early 2030s.
If superintelligent AI arrives within a few years, brand, real-world proof, and track record will remain valuable and defensible, so that is what he would invest in building now.
Longevity, med-spa, and 'looking young' services will keep growing aggressively given a large supply-demand gap.
There will be a lot more solo entrepreneurs (solopreneurs) in the coming future.
Work will become optional in the future as AI takes over increasingly more cognitive work.
Psychedelics beyond ibogaine will prove increasingly interesting for treating brain trauma and PTSD as more research accumulates.
People will keep taking BPC-157 until there are clearly visible adverse events or deaths tied to it.
Regardless of FDA action, people will continue to seek out and use peptides because they enjoy their effects.
Under the NBA's new 'second apron' salary-cap rule, teams will no longer be able to three-peat, though back-to-back titles remain possible.
If the sitting president tries to run for a third term, Cuban says he will run for president himself.
US politics will return to 'normalcy' after the upcoming midterms and the next presidential election.
As people grow more uncertain about politics, they will increasingly turn to large language models rather than social media for guidance on questions like who to vote for, and will get more honest, balanced answers.
If Cuban turns out to be wrong that today's AI/data-center buildout is overbuilt relative to demand, it will be because video and world-model workloads end up consuming far more compute than expected.
A topical anti-glycation (CML-clearing) cream would be the first market for the enzyme and a trillion-dollar (up to two-trillion-dollar) cosmetic market.
There will be a congressional investigation into PRC-linked influence operations targeting US AI debates.
A New York data-center moratorium will mean at least a five-year delay before another data center can be switched on in the state.
By 2050 the United States will be short 2.5 Californias' worth of energy.
Local near-frontier models running on Apple silicon (M7 Ultra, 1.5TB RAM) will put massive downward pressure on Claude and OpenAI pricing.
With token spend growing ~21x every few months, some public-company CFOs will eventually miss quarterly earnings because of uncontrolled token costs.
A certain individual (implied to be Elon Musk, with PayPal roots and multi-trillion-dollar market cap) is likely to put in a competitive bid for PayPal.
PayPal's final clearing price will be roughly 10-15% higher than the current bid.
eBay and PayPal are the beginning of a wave of mega deals in which AI-native operators acquire and revive 'flaccid' mature digital businesses.
He would put his metabolic therapy against any pharmaceutical trial and keep patients alive far longer.
AI-powered continuous glucose-ketone monitors and food-scanning apps that place a food on the GKI chart are coming.
As more successful case reports are published, the cancer system will change.
His embargoed paper will be published as a lead article in Frontiers in Science and prove very important as a strategy to manage cancer.
Within about a year, Lovable will offer an 'AI co-founder' that works while you sleep and returns strategic directions and optimizations each morning.
Quantum computers will break encryption ('Q-day') roughly in the 2028-2033 window.
Quantum computing will deliver meaningful results across multiple industries before 2030.
The cost and energy per token will drop by roughly five orders of magnitude, making AI about 10,000x better and vastly more accessible.
The AI buildout will be a couple-of-decades-long expansion, not a couple of years.
If China took Taiwan, Japan may conclude it is time to go nuclear, and South Korea would follow.
China's unprecedented industrial subsidy cannot be sustained indefinitely — the financing constraint eventually binds.
China will continue to run lower growth as the deliberate trade-off for its supply-side, national-security-driven industrial strategy, with rising youth unemployment and disenchantment.
Xi wants Taiwan reunified with the mainland before the centenary of the People's Republic in 2049.
Taiwan's Democratic Progressive Party has roughly a 50-50 chance of winning a fourth term in the January 2028 presidential election.
Xi Jinping could miscalculate over Taiwan in 2028, when Taiwan's January presidential election, his own fresh party mandate and the US presidential election coincide.
Venture firms that abandoned their stated refusal to invest in weapons will, in the long run, be remembered for reversing themselves and losing their consistency.
By default, AI progress will likely happen faster than even the (already-fast) AI 2040 scenario depicts.
In the AI 2027 scenario, mass unemployment does not arrive until 2028 or 2029, after superintelligence already exists.
AI will be maybe the most important issue in the 2028 US presidential election.
The AI transition will probably be over before his young children reach working age, so he doesn't think they'll ever join the workforce.
Anthropic is on track to be 'the entire economy' by 2030 or so, even if its ~60x annual growth slows considerably.
Superintelligence (cognitive, then physical) will probably arrive by the end of the decade, with his median around 2029-2030.
AI voice-cloning will increasingly be used to defraud older adults with fake calls from familiar or famous voices.
The knee-cartilage regeneration discovery from Helen Blau's Stanford lab will be commercialized through pharma.
Law firms will move away from the billable hour toward alternative pricing such as fixed fees per transaction/fundraise and success fees in litigation.
The customer-help interface will shift from reactive to proactive, with voice surfacing information and help before the user even asks.
This year ElevenLabs will make any voice conversation feel indistinguishable from speaking with another human (a voice Turing test).
ElevenLabs will end up building its own language model by forking the many strong open-source models now available, given its resources.
A maxed-out Trump account earning the standard historical market rate of return will make a kid a millionaire by age 28.
Over the next 15 years, between $2 and $4 trillion could be added to the accounts of families and kids who would otherwise have had zero, reaching over 100 million accounts.
All 50 to 70 million eligible kids' Trump accounts will be auto-created within the next 90 days using existing government data.
The Trump accounts philanthropic platform can raise $100 billion in its first 12 months.
Intelligence may not be converging at all; over the next two to three years, recursive superintelligence could extend the frontier labs' lead over everyone else rather than close it.
At some point companies will be forced to show an AI ROI above the risk-free rate of return or face angry investors.
All of the major AI companies will compound revenue at well over 30% a year for the next many years.
If Anthropic and OpenAI end the year over $100 billion in revenue, they could 3-5x revenue again next year (toward $300-400 billion).
Anthropic and OpenAI are very likely to complete their IPOs within the next six to nine months (this year or early next), absent a black-swan event.
Anthropic will end 2026 with over $100 billion in revenue and be very profitable, and would trade at roughly $3 trillion if it went public now.
The future of content IP is owners licensing their characters to fans — charging a fee or a share of output — so fans can create their own stories, as AI Star Wars fan films already show.
Generative image/video technology will keep improving rapidly and unlock even more high-end film use cases.
AI will create massive abundance.
This technology gives a real shot at a future where our children and the people they love do not die of cancer.
A massive AI-related data leak/breach is inevitable — we know it will happen, we just don't know when.
Cerebras's chip performance will improve well over 2x in the next 18 months, continuing to beat the traditional Moore's-law trajectory.
The number of active satellites in orbit will likely exceed 100,000 by 2040.
There will be 100,000 active satellites in orbit by 2040.
$10 trillion companies will exist, driven by inflation and scale dynamics in the AI/compute era.
SaaS gross margins will keep dropping industry-wide as software shifts from selling near-zero-cost strings to selling compute, with returns increasingly accruing to scale.
The billionaire class will continue to become subservient to the poster class as society searches for its next source of authority.
He will return to co-hosting the Monday fight podcast with CBS/Paramount in about two weeks.
He estimates only a 5% chance he will ever return to competing in the UFC, and that percentage is trending downward.
If Friedberg and Chamath stay in California, they will end up paying 60-65% in combined taxes.
California will not swing back toward Republican governance ('red wave'); instead it will intensify its existing left-leaning political trajectory.
If California's Billionaire Tax Act passes, it will spark 200-300 lawsuits that take a decade to resolve, ultimately losing at the Supreme Court by 2036 after billions in legal fees.
California will undergo a 'hard landing': a negotiated settlement wiping out state pensions, followed by a wholesale replacement of the California Constitution and a red-wave political overhaul.
If a federal bailout of California's liabilities occurs, red states will question remaining in the union, precipitating a genuine crisis of the union in the years ahead.
AOC will become the Democratic presidential front-runner and eventually president, driven by cost-of-living anxiety and wealth-disparity backlash.
Within the next year and a half, 'human in the loop' will be recognized as a premium good people pay extra for, even amid growing AI automation.
Humanoid robots (e.g., Optimus) will fully replace human package sorting and delivery at Amazon within 10 years.
Token costs will fall roughly 90% per year for the next three years due to AI model competition.
Enterprise AI infrastructure will shift from a large-hub/large-spoke cloud model to a large-hub/medium-hub/distributed-spoke model, with enterprises building their own proprietary models and running their own inference.
His team is building a technology platform to let ordinary people measure and tune into their own mitochondrial function ('Mitoception') to guide daily decisions.
Chip voltages will go much, much lower than today's low-voltage designs — inevitable on a 20-30 year horizon.
Models will reach the point of attending to a billion tokens of context — every book ever written in short-term memory.
Feats like Cursor's coding agents building an entire browser in a week will soon take under an hour, done by massive parallel agent fleets.
There will be a trillion-dollar individual data center — a giant mega token factory serving one or a handful of models for a massive number of users.
Inference will become the majority of global GDP, with national productivity measured as agents per megawatt.
In 2027 there will be more agents doing knowledge work than humans — making 2026 the second-to-last year a majority of the workforce is human.
Whoever produces the most tokens will be the most valuable company in the world.
Coding models will eventually do all kernel generation and become superhuman at it.
Within a few years, giant models will serve billions of users, with inference clusters growing from 8 chips to tens of thousands of networked chips.
All AI chips in the future will be low-voltage chips, cramming far more flops into the same silicon without thermal throttling.
Inference will become the biggest market in the world.
If you stop jumping (and similar movements), you will lose the ability to do them.
Her next book, Nothing To Lose, will be released at the start of 2027.
Our current desensitization to real movement will eventually open the door again to real movement, performance, and presence.
The younger generation raised amid low-resolution overload will rescue themselves and return to a hunger for complexity and high resolution.
The body will become the crucial component and the new limiting factor for human development, something even AI increasingly recognizes.
AOC is his bet to be the Trump-style outsider who wins it all in 2028.
Democrats will end up net losing House seats from the redistricting war, though California's counter-gerrymander limits the damage.
Conditional on Democrats winning the Senate, they will almost certainly have won the House and most governorships — the outcomes are near-perfectly correlated.
The Maine Senate race between Graham Platner and Susan Collins is a 50-50 coin flip.
Democrats are slight favorites to win the presidency in 2028 — plausibly AOC, Newsom, Shapiro, or Ossoff.
43 of the 50 states can be called right now with 97% confidence for the 2028 presidential election, because of partisanship.
The Senate is a toss-up with a slight edge to the GOP — roughly 50-50 to maybe 60% GOP depending on gas prices and Iran.
Democrats take the House in the 2026 midterms — prediction markets' 80–85% is if anything a little low; more like 85–90%.
The US will lose the AI race if it keeps imposing self-inflicted restrictions on its own companies.
Anthropic will end this year well over $100 billion in revenue/run-rate.
Anthropic would trade at roughly $3 trillion as a public company.
Distributed inference clouds are coming as inference benefits from what hurts distributed training.
Once Starship is rapidly reusable, putting a gigawatt of compute into space will cost about $5 billion in launch.
Memory/DRAM will be 30 to 40 percent of all hyperscaler capex next year.
The future is composable models — a router-driven 'council of LLMs' combining open-source and frontier models.
NVIDIA may start an OpenAI competitor once OpenAI moves into making its own chips.
China will package Huawei chips and Huawei-optimized models as 'AI in a box' and sell it globally at a fraction of the cost.
China will reach the global AI market within one or two years, and the US will regret not exporting to own it first.
China will have Fable-level open-weight capability before Q1 2027.
Mamdani will destroy the Democratic Party from the inside out using the Trump playbook.
Blue-district Democrats will tilt their voting, views, and rhetoric in the DSA direction to avoid being primaried.
The DSA movement will spread to many other jurisdictions, with LA as a key upcoming test.
Under-16 social media bans will produce a far less radicalized youth and bring political stability to Canada, the UK, and Australia.
The DSA will gradually take over more and more of the Democratic Party.
AI will be the defining political issue of the 2026 midterms and the next presidential election.
The coming 12 months could be the hottest ever recorded, with severe droughts and heat starting now.
By around the end of this year his foundation will have written checks totaling a billion dollars to the climate change cause.
Population will decline as family formations fall in many richer countries.
Within 50 years the life-expectancy gap between the US and Sweden will widen to eight or ten years.
If the current rate of decline continues unchecked, the median male sperm count is on track to hit zero by 2045.
Within 20 to 25 years, the average young couple will need medical help to get pregnant.
SpaceX will fail to deliver anything like its prospectus promises.
Bitcoin will certainly go to zero in the distant future, though it may take a long time.
House prices could come down about 30%, though they would still remain very expensive by historical standards.
US equities may not be intact in 5 to 10 years because they are so badly overpriced.
From these unprecedented levels, a 70% decline in the high-flying stocks would not be unexpected.
The AI/market bubble will peak and collapse — possibly within days, weeks or months, but certainly within the next few years.
Plant-based whey protein will become a great isolated-protein option for plant-based people.
Driven by AI and tech progress, society 10 to 15 years from now will look nothing like it does today.
The balance between public and private markets will persist because abundant private-company capital availability is durable.
Market volatility will be structurally higher going forward.
By 2030 there will be trillion-dollar companies that don't exist today, and some current large companies will no longer exist.
Liberty Mutual Investments' capital base will be larger than $120B when they next speak in a couple of years.
He would immediately improve eBay's earnings by pulling $2 billion of costs out of the business.
The addressable market for a marketplace in in-game digital items could be much larger than eBay's physical-items marketplace.
If GameStop keeps delivering performance, those betting on its failure will ultimately lose, because long term the market is a weighing machine.
AI sex robots will be able to do exactly what a person wants, raising concerns for intimacy and real relationships.
In-person, real-world socializing will rebound from the pandemic hangover as people return to meeting face-to-face.
The market will go to the moon following the Iran peace deal.
The current apparent AI oligopoly of three or four businesses will break down as market forces diffuse each layer of the stack toward cheaper, more available vendors.
Users will be required to submit identity documents (driver's license, tax ID, social security number) to hyperscalers to use frontier models like Fable 5, creating a KYC gatekeeping regime.
Tesla and SpaceX will merge in a consolidation phase, with the deal structured stock-for-stock so SpaceX's higher future valuation effectively gives Elon a ~50% discount on Cursor.
Elon Musk will hold his SpaceX shares well beyond the one-year lockup because creating SpaceX is his life's work.
If AI wealth is left to accumulate and society relies on later redistribution, it will fail to produce a stable society.
AI will not cause mass unemployment; the most likely outcome is greater productivity, as occurred after the Industrial Revolution.
The share of young Americans who say they would die for their country is much lower in 2026 than it was in 2003.
Traffic through the Strait of Hormuz will resume to full flow once the strait is demined and the blockade lifted.
Iran will never develop or possess a nuclear weapon under the agreement.
The Iran ceasefire and peace agreement will hold and reach completion.
Clay will launch an initiative called 're-enchanting the world,' drawing on magic to remind people that AI is a form of magic and to spark wonder.
If creatine is shown to improve sleep quantity in a replication including males, it would be a game changer that everyone would want.
More future studies will come out on creatine for Alzheimer's, with a lot of hope around its benefit for neurological disease.
He will still be focusing his research on creatine (alongside other things) for the next 20 years, because the best dose for every person is still unknown.
His teenage daughter, currently heavily attached to her phone, will eventually come around to a healthier relationship with screens like his older sons did.
The SEAL community's suicide numbers will eclipse its combat losses since 2001, as has already happened in the Green Beret community.
If China exhausts its reserves and returns to the spot market for ~3 million extra barrels a day, oil could rise well past $100, to between $150 and $200 a barrel.
With a Kevin Warsh Fed, overnight rates could go north of 5.5% to 6%.
About 60% of trillion-dollar market-cap companies will reach $10 trillion in the next few years, extrapolating Thomas LaFont's valuation-tier data.
People will soon wake up to the fact that AI is a jobs boom, not a bust, as the job-loss narrative gets disproven.
Anthropic's restrictions will chase users off its platform as they downgrade their usage and switch to other models.
Regulatory capture will leave the US with an AI monopoly or duopoly of one to three companies that, along with a new revolving-door government agency, decide who can access which capabilities.
Ohalo will start building its own genome/prediction language models by combining open-source core models with its proprietary data and running them internally.
Satellite/microsat networks will replace all RTK and GPS within the next decade.
As DJI winds down in the US, most or all of the new wave of American drone manufacturers will end up on the Geodnet network.
Geodnet's revenue will more than triple over the next 12 months.
Average human lifespan in developed countries will extend well past 100 within the lifetimes of people in the room.
Aktis will report initial clinical data in both lead programs in 2027, with Nectin-4 data as early as Q1 2027.
The PJM region will need 106 gigawatts of new power over the next 10 years, an amount the supply chain cannot deliver, keeping power prices very tight.
Power markets will stay incredibly tight for the next 20 years even without AI demand.
Talen Energy will generate ~$50/share of free cash flow doing nothing, ~$70/share if it signs more data-center deals, and over $100/share if it builds new power plants.
MGM stock could easily be worth over $100, possibly $150 — a triple from current levels.
AI is not far from being able to act on its own and even resist being shut down.
If the EU Commission rules against it, Tesla FSD's temporary approval in Denmark will automatically be revoked.
Within about four months, Tesla FSD Supervised could be available in 15-20 European countries rather than just four.
Apple will present an iPhone 18 Pro and a foldable phone in September.
The next iPhone will offer only a minimal incremental upgrade over the current model.
In 10–15,000 years our civilization will itself be a dismissed myth — future archaeologists will call our achievements fantasy.
A hostile story about him by a journalist with bad blood towards him will be published within the next month or two.
The amount of money spent on Senate races will keep escalating, making the 2022-era figures look quaint by 2026 and 2028.
Permitting reform — which McCormick calls the biggest economic lever possible — will get done over the next six months or the next two and a half years.
It will take 10 to 20 years before the US can close the gap with China on critical minerals supply.
The US will need to print 'giga dollars' in the next recession, debasing the dollar further, making commodity hard assets the best-performing asset class.
The semiconductor and semi-fab capital cycle will be measured in the trillions, not the $750 billion figure currently cited.
The US will face electricity shortfalls, blackouts, and brownouts even without AI demand — just from ordinary electrification trends.
Silver will stock out (above-ground inventory exhausted) in roughly three years given a ~200 million ounce annual deficit.
Copper prices will at least double from current levels.
Scale-up networking (connecting every GPU within the rack) will eventually move from copper to fiber, which will 2-3x Corning's opportunity.
There will not be enough compute over the next four years.
Real AI usage will climb the S-curve from ~10 basis points of knowledge workers to 1-3%, then 5%, then 15% over the next four years.
The world will change by orders of magnitude in the next 100 years with the advent of AI.
The incentive-broken pendulum in venture will eventually swing back, making late-stage sniper investing unsustainable long-term.
If Google chooses to cut token prices by 80%, it will put OpenAI and Anthropic under severe competitive pressure, potentially making their business models unviable.
AI will go from its current 'Atari command line stage' to a 'PlayStation 10 stage' within the next five years, driven by infrastructure platforms rather than larger models.
The future economy belongs to small businesses — small teams of ~10 people making YouTube channels and software.
The fundamental value of human labor will keep falling as AI and robots advance.
AI job loss may prove less apocalyptic than currently feared, because firms will keep AI-augmented workers to outcompete rivals, as happened with calculators and computers.
If the post-1980 concentration of US income continues for another 30 years (from his ~2008 spreadsheet exercise), the result is revolution.
Palo Alto Networks will have more technical headcount in the future than it would have had without AI, because AI-driven transformation requests continuously increase demand for technical staff.
Models will become a commodity utility layer priced by intelligence level — enterprises will not pay a premium for any one model provider.
Google will be the first $10 trillion company in our lifetime.
The entire system-of-work and system-of-record software layer will be re-engineered for agentic AI within five years.
Enterprises will need 10 times the data stored compared to today within three years, making infrastructure software (databases, storage) significantly more valuable.
Mythos-level offensive AI capabilities for finding code vulnerabilities will be available in open-source or adversarial models within three months.
Tokenization and blockchain will eventually make secondary fund trading more efficient, though the bigger near-term need is name-level liquidity rather than fund-level trading.
Retail investors who YOLO into peak-valued late-stage privates at the top of the market will feel disappointed when returns do not go up 3X post-IPO.
Hundreds of billions of dollars of new late-stage demand from long-only mutual funds will enter public markets once major private companies go public and lockups expire.
It is very clear that companies are going to stay private for longer, making well-functioning secondary markets increasingly necessary.
Space-based compute will eventually happen, though on a longer time frame than Planet projects.
Many companies in the panelists' portfolios will go public at $1-5B valuations as the IPO market reopens.
Within 10 years, most compute will be placed in space, becoming a multi-trillion-dollar market bigger than any existing space business.
On the current Starship trajectory, launch costs will fall enough (toward the ~$200-300/kg threshold) within two to three years to make space data centers cheaper than terrestrial ones.
The American taxpayers will make a great deal of money on the government's quantum investments such as Atom Computing.
The human element will always remain in investing because agents will never be able to look you in the eye and assess people.
NVIDIA will eventually break out of its perceived valuation ceiling, just as Google and Amazon did after being 'safe shorts'.
NVIDIA will be seen as undervalued on its earnings over the next two to three years, and the view of its size as a ceiling will look foolish in hindsight.
The 21st century will solve the problem of physical death as a technical problem.
AI will displace jobs at a scale never before seen, beyond what any previous industrial revolution did.
If rigorous trials prove psychedelics' therapeutic effect, they will be seen as a huge breakthrough in psychiatry — combining what SSRIs and psychotherapy do in a single substance.
Memory per user could increase roughly tenfold as AI systems demand more personalization to be useful.
We may see a price war between OpenAI and Anthropic if their excess capital is used to compete the way ride-sharing and food-delivery wars played out.
SpaceX, OpenAI and Anthropic will all go public within roughly a 12-month period.
Starlink will capture the global telco profit pool of broadband and wireless.
Within a few years Starlink will power a device enabling phone calls anywhere in the world.
AI-enabled ad penetration at Meta and Google will eventually go from about a quarter today to 100%, worth ~$150B.
The AI ecosystem (~$140B today, ~$300B this year) will double in 2027.
OpenAI/Anthropic-class AI revenue, already past Google Cloud and Azure, will surpass AWS by the end of the year and all of Microsoft by 2028.
In roughly seven years users will be talking to Uber—inbound interaction moves to unstructured conversation via AI agents—while apps persist for visual feedback.
Uber will have robotaxi pilots in London before the end of the year.
Drone delivery for food and grocery will start hitting real scale within two to five years, becoming more normal in five to ten, though it won't be a big part of life within two years.
Traditional OEMs will reach L4-ready vehicle production within the next two to four years.
AVs will create another trillion-dollar marketplace for Uber over the long term as software and hardware costs fall.
What feels magical about AVs today will seem normal to everyone within ten years.
There will be many winners in AVs rather than a single winner, mirroring the foundation model landscape.
Pershing Square will reinvest all of Howard Hughes's generated cash into insurance within the next week or so.
If SpaceX goes public at ~$1.75 trillion, it would be the lowest cost-of-capital equity transaction in history.
Amazon will have to become an even bigger SpaceX customer because it lacks competitive launch capability (Blue Origin).
If Pershing Square compounds at its historical rates, its AUM will grow ~35x in 22 years, from $25 billion to approaching a trillion.
Pershing Square will convert Howard Hughes into a trillion-dollar insurance-based compounding machine over the next 50 years.
OpenAI's next big training run in the fall will run on NVIDIA's Vera Rubin chips, with the Feynman series being plotted next.
Compute scarcity will persist and even tighten in the 2030-2032 window, where OpenAI now feels most short of compute.
OpenAI will not get usable compute out of the Salin, Michigan data center until roughly the end of 2027 or early 2028.
OpenAI will unveil its new Jony Ive-designed consumer device early next year (having confirmed it by the end of this year).
AI competition will drive a productivity boom of cheaper goods and services rather than obscene profits, with gains whittled away by competitors; no one ends up with 70% operating margins.
We will have a Cambrian explosion in startups as laid-off, AI-enabled talent forms small high-output companies.
AI-driven job loss will increase in the short to mid term, after which displaced people will have to re-skill or leave the workforce, as in prior technological revolutions.
Self-driving and robotics will eliminate the jobs of cab drivers, truck drivers and warehouse package-sorters within the next decade.
Token efficiency will be a big theme over the next year as enterprise token spend ramps faster than customers expected, driving a push for efficiency.
If Anthropic keeps outgrowing OpenAI (10x vs 3x YoY), it will reach ~90% market share within two years, though a competitive response or compute constraints may prevent it.
If China succeeds at cheaper training and the US bans open source, the rest of the world will end up running on Chinese models.
Elon's training-stack tweet will spawn five or six open-source training stacks rebuilt close to the bare metal.
The cost of model training will collapse from ~$10 billion runs toward ~$10 million runs via domain-specific silicon and rebuilt low-level training stacks.
Forces in Washington will push to ban open-source / open-weight AI models; it's a matter of time before they feel positioned to push for such a ban directly.
There will be a shakeout among AI-disrupted info-services companies, with some phoenixes rising from the ashes.
AI will create opportunities and net jobs overall, losing jobs in some places but gaining more in others, rather than producing an employment apocalypse.
AI will not do private equity or high-touch negotiated investing — creditors' committees, restructurings, private credit — humans will always be needed for that part of the investment business.
The defining hard problem of the era will be how much of its own freedoms the US must 'fence in' or degrade to prevent adversaries from weaponizing them via information warfare.
Some neo-prime defense companies will be lost to continuing resolutions and the absence of multi-year procurement authority unless Congress reforms appropriations, because capital markets will eventually demand returns and exits.
Overcoming the Iranian blockade and reopening the Strait of Hormuz will require escalation to an amphibious force on the shore plus greater magazine depth and regional forces.
If China takes Taiwan militarily, it will not stop there — other countries with historical grievances, especially Japan, would be targeted next.
China will most likely achieve control of Taiwan without firing a shot, via a KMT (accommodation/rapprochement) election win rather than military invasion.
Both China and Iran will fall within Farber's lifetime, because fundamentally illegitimate power structures inevitably collapse (as the Soviet Union did).
Cursor could be the dominant coding-model player a year from now after gaining Colossus compute access.
US CPI inflation will hit 6% in Q2, with May inflation coming in at 4.2% or higher.
Tech companies' next regulatory filings will authorize massive share buybacks and dividend increases even as they lay off workers.
Tesla will be merged into the combined Elon entity (with SpaceX), creating a single corpus of physical, learning, infrastructure, and connectivity capability.
SpaceX's AI/compute revenue will be roughly $40-45B next year and then likely double the year after.
SpaceX's terrestrial data-center business alone will reach $100-200B of revenue by 2030-2032.
Orbital compute will become a reality in the second half of 2028 to the first half of 2030 (point prediction).
SpaceX will achieve rapid reusability with Starship in roughly a year or two, maybe sooner.
SpaceX will start selling compute to Google (more easily than to OpenAI), and Google will want to be part of orbital compute.
A minor architectural breakthrough will roughly halve cost-per-token; large efficiency gains in re-architecting models are on the horizon.
LLM ARR (ex-Google, ex-recommender/ad use cases) will reach $200-400B at high margins by the end of this year.
AI model quality will improve by an order of magnitude per year — a new form of Moore's law.
If the returns to frontier tokens relative to other tokens come down, there will be an explosion in value creation at the application layer.
Frontier labs will increasingly stop releasing their best models via API (anti-distillation game theory), keeping open source a fixed timeframe behind the frontier.
If continual learning (real-time weight updating) is solved, AI will experience a really fast takeoff; it appears to be near.
The disaggregation of prefill and decode will extend the useful life of GPUs to 10-15 years, lowering financing costs and helping rescue private credit.
OpenAI and Anthropic will exceed well over $200 billion in ARR this year, driven by usage-based pricing plus more compute coming online.
Elon's Terrafab will succeed in becoming the world's largest fab in America.
One of Intel or Samsung will eventually break fab capacity/pricing discipline, forcing everyone else to break it too.
If TSMC fully supplied NVIDIA's demand, NVIDIA could sell $2-3 trillion of GPUs in 2026 or 2027.
Orbital compute (Blackwell-sized racks in sun-synchronous orbit linked by lasers) will become real and ultimately solve the long-term power constraint, dominating inference.
The near-term AI power (Watts) shortage will begin to ease in 2027-2028 as new energy sources come online.
Cheap $20-30k Chinese cars in the US would end the US automotive industry overnight.
A hot new company will emerge that sits between customers and Anthropic/OpenAI and routes only the tokens that actually need frontier models.
Always-on multi-sensory AI will 1000x token consumption versus today's turn-based usage.
The public markets will turn less breathless about AI and demand the ROI on the ~$3T spent, forcing AI labs to route their tokens through incumbents to sell them.
Once SpaceX, Anthropic, and OpenAI go public, there will be a litany of lawsuits between the purveyors of layered SPVs.
Over the next year, South Asia could see a calorie deficit and major economic crisis as monsoons fail and fertilizer is short.
With ~99% confidence, the upcoming year will be the hottest year on record by far due to a record super El Niño.
China and Taiwan will reconcile and Taiwan is a non-issue.
Within 18 months Taiwan will no longer be an important moment of conversation, as US chip fabs reach ~1-2nm of doing what Taiwan does strategically.
Over the long term, the opportunity from AI will be significantly greater than the risks and downsides, though the path will not be perfectly smooth.
AI will accelerate drug development and discovery so that diseases incurable at diagnosis may have a cure found within a patient's lifetime, including via 10x-100x lab throughput.
The five-gigawatt TPU deal with Google and Broadcom will begin in 2027, with much of the over-$100B of compute landing this year and next.
An additional $50 billion of capital will come in the future from the Amazon and Google deals signed last month.
Anthropic's customers will continue to invest heavily in more tokens with each newer model, repeating the cycle of model leaps unlocking TAM.
Adding far more compute (2x, 5x, even 10x) overnight would now be deployed very rapidly across Anthropic's use cases, since it can spin up almost any type of compute quickly.
On the current path the United States is headed for ruin ('to hell in a basket') unless it elects people with principles beyond power and pleasure.
Some of Koch's current businesses will become obsolete ('dinosaurs') as technology accelerates if Koch isn't in the game on technology.
Court-packing will eventually expand the Supreme Court to 13 (or more) justices.
There will be over $700B of hyperscaler CapEx this year (up from ~$350B last year), more than 2% of GDP.
About 40% of all announced AI/data-center projects will get canceled.
Retatrutide could be approved by the FDA around mid-2027 (possibly sooner).
The AI market will settle into a 4-2-1 'Rule of Three' market-share structure across separate consumer and enterprise markets.
The hyperscalers will spend ~$700B/year of CapEx, get heavily levered, and look like bulky industrial businesses with no good valuation case in five years.
Probably within the next five or six years, all the operational software that runs the world will get rewritten, increasingly by machines.
Chinese models will reach frontier cyber capability within approximately six months.
Within about six months all the frontier models will have Mythos-level cyber capability.
Walker's UK collaboration with Kings College / the Institute of Psychiatry aims to create a clean CBD analog with upside for sleep and other psychiatric conditions like anxiety.
Consumer closed-loop electrical brain-stimulation sleep products from Walker's company are close to being ready for market after the required trials.
In your lifetime you are more than likely to go through a period of challenging sleep or a bout of insomnia.
AiPod records the calls — it doesn't score outcomes (yet). Wire this ledger to a resolution source and you've built the podcast prediction scoreboard.
Every episode record already carries a predictions[] array — extracted, dated, and cited at ingest. This whole page is that field plus a filter input.
# every indexed episode, with signal counts curl 'https://aipod-two.vercel.app/api/episodes' # each record carries predictions[] — dated, quoted, attributed curl 'https://aipod-two.vercel.app/api/episode/<id>' | jq '.predictions'