AI predictions this week: Google leads, bubble fears rise
AI predictions show rising concern about a bubble, lower Anthropic valuation expectations and stronger calls for Google and major math breakthroughs.
AI bubble predictions this week
The crowd’s AI outlook has split in a striking way. Concern about valuations has risen sharply, while expectations for a major mathematical breakthrough have also strengthened. These calls describe different outcomes: AI could become more capable even as expectations for company values fall.
The AI bubble question gives a burst by December 31, 2026 a 64% chance, up from a 9% chance a week earlier. That is a rise of 55 percentage points. The June 30, 2027 deadline has a 69% chance, up from a 20% chance.
Chances move all the time; we read these at 21 November 2026, 16:43 UTC. Our numbers give no stated reason for the week’s moves.
The important distinction is between direction and timing. Both bubble calls now lean Yes, but the later deadline sits only modestly higher. A useful next check is what the question counts as a burst: a broad label needs a clear outcome before readers can judge whether the crowd’s call makes sense.
Anthropic valuation forecast
The November Anthropic valuation calls show the clearest shift towards lower outcomes. The crowd has sharply raised the downside call and reduced its expectations for higher thresholds.
The December valuation question extends that pattern, but also shows its limits. The downside $1.0T call has a 65% chance. The downside $800B call has a 26% chance, up from a 9% chance, while the downside $700B call has a 17% chance, up from a 6% chance.
That is not a prediction that every lower threshold will be reached. The crowd favours the $1.0T downside outcome but still gives deeper declines much lower chances. On the upside, the $2.0T call has a 14% chance, down from a 26% chance.
These are threshold predictions, not statements of Anthropic’s current valuation. A confirmed valuation relevant to a question’s terms would be the clearest evidence to watch. Keep the November and December deadlines separate: they ask what happens over different periods.
- The downside $1.275T call for November 30 has a 97% chance, up from a 47% chance.
- The upside $1.575T call has a 9% chance, down from a 53% chance.
- The upside $1.6T call has an 8% chance, down from a 57% chance.
Can AI solve major math problems?
The strongest technical optimism centres on the Riemann Hypothesis call. It now has a 67% chance, up from a 9% chance. That 58-percentage-point rise stands in sharp contrast to the weaker valuation outlook.
The related Millennium Prize Problems question gives exactly one more solution in 2026 a 64% chance, up from a 10% chance. The zero-more outcome has a 26% chance, down from an 85% chance. The crowd has shifted from expecting no further solution towards expecting one.
The optimism does not spread evenly across mathematics. In the individual problem calls, Ramsey Number R(5,5) has a 25% chance, up from a 16% chance. The Hodge Conjecture has a 20% chance, up from a 15% chance. Both have strengthened, but neither is the crowd’s expected outcome.
What would move these predictions? A claimed solution would need serious mathematical scrutiny. The key distinction is between an impressive claim and a proof that withstands checking. Readers should also distinguish a call about a named problem from a call about the total number solved.
Who will have the best AI model?
Google leads the year-end AI model question. The crowd gives Google an 85% chance, up from a 74% chance. Anthropic has an 11% chance, down from a 22% chance.
The matching moves show a widening gap between these calls. Google has gained 11 percentage points while Anthropic has lost 11 percentage points. This is a much more decisive preference than the bubble or leading mathematics predictions.
The useful next question is what would challenge that lead. A model release could matter if its performance changes the comparison used to decide the outcome. A launch announcement alone would not establish leadership. Read the question’s criteria rather than treating “best” as a universal judgement.
Model leadership and valuation also measure different things. A company can improve its technology without reaching an upside valuation threshold.
Which AI calls are closest?
The calls with the most room for disagreement are not necessarily those with the biggest weekly moves. The December bubble prediction, the exactly-one-more Millennium outcome and the downside $1.0T Anthropic threshold all lean Yes, but none shows the overwhelming preference seen in the November $1.275T downside call.
Start with the outcome you want to assess: technical progress, company value or model leadership. Then compare its deadline and the evidence needed to settle it. Do not treat these questions as interchangeable votes on whether AI will succeed.
Browse the AI topic page, look at the wording behind the chance, and make your own call when you are ready.
Şu an canlı
Bu yazıdaki ihtimaller yazıldığı anda okundu. Bunlar canlı rakamlar.
AI bubble burst by June 30, 2027?
68%olasılıkSık sorulan sorular
- Does the crowd expect an AI bubble burst?
- The crowd gives an AI bubble burst by December 31, 2026 a 64% chance, up from a 9% chance a week earlier. The June 30, 2027 call has a 69% chance.
- Who is favoured to have the best AI model?
- Google has an 85% chance in the year-end model question, compared with an 11% chance for Anthropic. Google’s call rose 11 percentage points over the week.
- Will AI solve the Riemann Hypothesis in 2026?
- The crowd gives AI solving the Riemann Hypothesis in 2026 a 67% chance, up from a 9% chance. That is a prediction about the outcome, not confirmation of a solution.
Sitemizdeki canlı ihtimallerden yapay zeka yardımıyla yazıldı ve yayımlanmadan önce otomatik olarak kontrol edildi. İhtimaller kalabalığın görüşüdür, tavsiye değildir. Yalnızca 18 yaş ve üzeri. Fiyatlar ve sonuçlar nasıl işler.