Pending ●●●●●
The AI stock bubble will burst within about a year, i.e. by the end of September 2027 (moved up from his earlier 2028 base case).
“the bubble in AI may burst sooner than later.”AI
Who gets the future right?
A public record of predictions made by well-known people: their exact words, when they said them, and how things turned out.
Scorecard
| Person | Category | Predictions | Scored | Accurate | Late | Partial | Wrong | Accuracy |
|---|---|---|---|---|---|---|---|---|
| Al Gore | Politician | 8 | 5 | 1 | 0 | 0 | 4 | 20% |
| Allan Lichtman | 8 | 7 | 5 | 0 | 1 | 1 | 79% | |
| Anthony Fauci | Scientist | 9 | 5 | 5 | 0 | 0 | 0 | 100% |
| Bill Gates | Tech leader | 14 | 5 | 2 | 0 | 0 | 3 | 40% |
| Cathie Wood | Investor | 11 | 6 | 3 | 0 | 0 | 3 | 50% |
| Dario Amodei | Tech leader | 8 | 2 | 0 | 1 | 1 | 0 | 50% |
| Donald Trump | Politician | 12 | 7 | 2 | 0 | 0 | 5 | 29% |
| Ed Yardeni | Market strategist | 9 | 5 | 4 | 0 | 0 | 1 | 80% |
| Elon Musk | Tech leader | 8 | 5 | 0 | 2 | 0 | 3 | 20% |
| Eric Schmidt | Tech leader | 10 | 5 | 1 | 0 | 0 | 4 | 20% |
| Geoffrey Hinton | AI researcher | 7 | 1 | 0 | 0 | 0 | 1 | 0% |
| Gordon Chang | Commentator | 3 | 3 | 1 | 0 | 0 | 2 | 33% |
| Harry Dent | Market strategist | 10 | 9 | 1 | 0 | 0 | 8 | 11% |
| Jack Dorsey | Tech leader | 5 | 1 | 0 | 0 | 0 | 1 | 0% |
| James Hansen | Scientist | 11 | 6 | 6 | 0 | 0 | 0 | 100% |
| Jamie Dimon | Business leader | 10 | 3 | 2 | 0 | 0 | 1 | 67% |
| Janet Yellen | Policymaker | 8 | 5 | 2 | 0 | 0 | 3 | 40% |
| Jensen Huang | Tech leader | 8 | 1 | 0 | 0 | 0 | 1 | 0% |
| Jeremy Siegel | Economist | 11 | 10 | 4 | 1 | 0 | 5 | 45% |
| Jerome Powell | Policymaker | 11 | 9 | 4 | 0 | 4 | 1 | 67% |
| Jim Cramer | Commentator | 8 | 7 | 1 | 0 | 0 | 6 | 14% |
| Joe Biden | Politician | 10 | 7 | 5 | 0 | 0 | 2 | 71% |
| Kevin Hassett | Economist | 7 | 5 | 1 | 0 | 0 | 4 | 20% |
| Larry Summers | Economist | 12 | 9 | 4 | 0 | 0 | 5 | 44% |
| Marc Andreessen | Investor | 8 | 5 | 3 | 1 | 1 | 0 | 80% |
| Mark Zuckerberg | Tech leader | 11 | 2 | 1 | 0 | 0 | 1 | 50% |
| Masayoshi Son | Investor | 5 | 0 | 0 | 0 | 0 | 0 | — |
| Michael Burry | Investor | 7 | 3 | 0 | 0 | 1 | 2 | 17% |
| Michael Saylor | Crypto figure | 10 | 4 | 1 | 0 | 0 | 3 | 25% |
| Michio Kaku | Scientist | 8 | 2 | 0 | 0 | 1 | 1 | 25% |
| Mohamed El-Erian | Economist | 9 | 6 | 2 | 0 | 2 | 2 | 50% |
| Nouriel Roubini | Economist | 10 | 7 | 2 | 0 | 1 | 4 | 36% |
| Paul Ehrlich | Scientist | 9 | 5 | 0 | 0 | 0 | 5 | 0% |
| Paul Krugman | Economist | 8 | 7 | 4 | 0 | 0 | 3 | 57% |
| Peter Schiff | Investor | 13 | 7 | 3 | 1 | 0 | 3 | 50% |
| Peter Thiel | Investor | 3 | 3 | 0 | 0 | 1 | 2 | 17% |
| Peter Zeihan | Commentator | 9 | 4 | 1 | 0 | 1 | 2 | 38% |
| Ray Kurzweil | Futurist | 8 | 4 | 1 | 1 | 0 | 2 | 38% |
| Robert Kiyosaki | Commentator | 9 | 7 | 1 | 1 | 0 | 5 | 21% |
| Sam Altman | Tech leader | 10 | 2 | 2 | 0 | 0 | 0 | 100% |
| Steve Ballmer | Tech leader | 6 | 4 | 0 | 0 | 0 | 4 | 0% |
| Thomas Friedman | Commentator | 5 | 5 | 4 | 0 | 0 | 1 | 80% |
| Tom Lee | 10 | 9 | 3 | 0 | 1 | 5 | 39% | |
| Vinod Khosla | Investor | 7 | 1 | 0 | 0 | 0 | 1 | 0% |
| Vitalik Buterin | 8 | 5 | 3 | 1 | 0 | 1 | 70% | |
| Yann LeCun | AI researcher | 6 | 1 | 0 | 0 | 0 | 1 | 0% |
Accuracy counts accurate as 1, late or partial as ½, inaccurate as 0. Pending predictions don't count.
Predictions
Pending ●●●●●
There is a decent chance "powerful AI" (smarter than a Nobel Prize winner across most fields, working autonomously) arrives within 1-2 years (by early 2028), and a very strong chance it arrives within the next few years.
“not that powerful AI is definitely coming in 1–2 years, but that there’s a decent chance it does, and a very strong chance it comes in the next few.”AI
Pending ●●●●●
By 2030, AI models will be extraordinarily capable and able to do things humans themselves cannot do.
“by the end of this decade, by 2030, if we don't have extraordinarily capable models that do things that we ourselves cannot do, I'd be very surprised”AI
Pending ●●●●●
Spending on AI infrastructure will reach $3 trillion to $4 trillion (per year) by 2030.
“We see $3 trillion to $4 trillion in AI infrastructure spend by the end of the decade.”AI
Pending ●●●●●
Superhuman AI of the kind described in the AI 2027 scenario will not arrive by the end of 2027.
“I acknowledge that I personally have longer-than-2027 timelines”AI
Pending ●●●●●
By mid-2030, AI will be able to do 80% of the work of any economically valuable job humans do.
“Within the next five years, any economically valuable job humans can do, AI will be able to do 80% of it”AI
Pending ●●●●●
In 2026, AI systems capable of producing genuinely novel insights will arrive.
“2026 will likely see the arrival of systems that can figure out novel insights.”AI
Pending ●●●●●
By the end of 2027, robots that can do useful tasks in the real world will arrive.
“2027 may see the arrival of robots that can do tasks in the real world.”AI
Pending ●●●●●
By about April 2026, roughly half of Meta's (Llama) software development will be done by AI rather than people.
“our bet is sort of that in the next year... maybe half the development is going to be done by AI as opposed to people”AI
Pending ●●●●●
By late October 2026, most of the code for Meta's AI (Llama) research efforts will be written by AI.
“I would guess that sometime in the next 12 to 18 months, we'll reach the point where most of the code that's going toward these efforts is written by AI.”AI
Pending ●●●●●
AGI/superintelligence (AI surpassing humans across domains) will arrive by April 2035.
“there's a good chance it'll be here in 10 years or less now.”AI
Inaccurate ●●●●●
By April 2026, the vast majority of human programmers will have been replaced by AI programmers.
“So we believe as an industry that in the next one year the vast majority of programmers will be replaced by AI programmers.”AI
Pending ●●●●●
By April 2026, AI systems will perform at the level of top graduate mathematics students.
“We also believe that within one year you will have graduate level mathematicians that are at the tippy top of graduate math programs.”AI
Pending ●●●●●
Artificial general intelligence, as smart as the smartest human experts across fields, will exist by April 2030.
“within three to five years we'll have what is called general intelligence, AGI”AI
Pending ●●●●●
Humanoid robots will be in widespread use (starting in factories) within a few years, before 2030.
“when, literally, humanoid robots are wandering around, which is not five years away. This is not five-years-away problem, this is a few-years-away problem.”AI
Partly right ●●●●●
AI will be writing 90% of all code within three to six months (by September 2025), and may be writing essentially all code within twelve months.
“we are not far from the world—I think we’ll be there in three to six months—where AI is writing 90 percent of the code. And then in twelve months, we may be in a world where AI is writing essentially all of the code.”AI
Pending ●●●●●
AI systems with capabilities akin to a "country of geniuses in a datacenter" will exist possibly by 2026-2027 and almost certainly no later than 2030.
“Possibly by 2026 or 2027 (and almost certainly no later than 2030), the capabilities of AI systems will be best thought of as akin to an entirely new state populated by highly intelligent people appearing on the global stage—a "country of geniuses in a datacenter"”AI
Accurate ●●●●●
Despite AI advances, people in 2025 will spend their time mostly the same way they did in 2024.
“Life will go on mostly the same in the short run, and people in 2025 will mostly spend their time in the same way they did in 2024.”AI
Pending ●●●●●
By the end of 2025, an OpenAI model will rank as the world's number one competitive programmer.
“I think our internal benchmark is now around 50, and maybe we'll hit number one by the end of this year.”AI
Pending ●●●●●
By early 2035, AI will make expert-quality medical advice and tutoring (the expertise of a great doctor or teacher) free and commonplace.
“with AI, over the next decade, that will become free, commonplace — great medical advice, great tutoring,”AI
Pending ●●●●●
Artificial general intelligence will arrive sooner than his previous "two or three years" forecast, i.e. by the end of 2027.
“But now I have corrected that statement and I feel that AGI will actually arrive much sooner than that”AI
Accurate ●●●●●
In 2025, a personalized AI assistant (expected to be Meta AI) will reach more than 1 billion people.
“In AI, I expect this is going to be the year when a highly intelligent and personalized AI assistant reaches more than 1 billion people, and I expect Meta AI to be that leading AI assistant.”AI
Pending ●●●●●
During 2025 it will become possible to build an AI engineering agent with the coding and problem-solving ability of a good mid-level software engineer.
“I also expect that 2025 will be the year when it becomes possible to build an AI engineering agent that has coding and problem-solving abilities of around a good mid-level engineer.”AI
Pending ●●●●●
By January 2030, current-style large language models will no longer be used (at least as the central component of AI systems).
“I think within five years, nobody in their right mind would use them anymore,”AI
Pending ●●●●●
By January 2030, a new AI architecture paradigm fundamentally different from today's LLMs will be in use.
“within three to five years we'll have systems that are a completely different paradigm.”AI
Pending ●●●●●
In 2025, the first AI agents will "join the workforce" and materially change companies' output.
“We believe that, in 2025, we may see the first AI agents "join the workforce" and materially change the output of companies.”AI
Pending ●●●●●
Human-level AI (AGI) will not be achieved within two years of late November 2024, i.e. not by end of November 2026.
“But it's not going to happen next year. It's not going to happen in two years,”AI
Right, but late ●●●●●
AI models will probably score about 90% on the SWE-bench software-engineering benchmark by November 2025, possibly sooner.
“And I think in another year we'll probably be at 90%. I mean, I don't know, but might even be less than that.”AI
Pending ●●●●●
Frontier AI training runs/clusters will cost a few billion dollars in 2025, possibly over $10 billion in 2026, and $100 billion clusters will probably be built by 2027.
“I think next year we're going to go to a few billion, and then 2026, we may go to above 10 billion. And probably by 2027, their ambitions to build hundred billion dollar clusters. And I think all of that actually will happen.”AI
Pending ●●●●●
Superintelligence will exist within a few thousand days of September 2024 (roughly by the early 2030s).
“It is possible that we will have superintelligence in a few thousand days (!); it may take longer, but I’m confident we’ll get there.”AI