When will it happen? We work it out, in public.

Several of the next cities lack a legal path to driverless commercial service, and those rules move on legislative calendars, not launch schedules.

On one reading, the launch pace settles this: four metros in September, ten needed in fifteen months. On the other, the cities still to come are not like the ones already served, because in several of them the law does not yet allow the service. New York dropped a proposal that would have allowed driverless taxis in c...

Waymo opened four metros in September alone, and its launch playbook is now routine; ten more in fifteen months is under its current pace.

Look at last month. Denver, San Diego and Tampa opened to public riders on 1 September, taking Waymo to 14 cities. Las Vegas followed two weeks later, on the same playbook it has used in 14 other cities: invite public riders on a rolling basis, then open up. That's four metros in one month. The company's stated goal w...

No magnetic-confinement device has yet measured Q above 1, and SPARC's dates have moved each year with no current completion date.

The design should be stated first. One first-of-a-kind device (n = 1), a compact high-field deuterium-tritium tokamak; primary endpoint, a measured Q above 1 in at least one discharge; confirming body, the operator, followed in practice by a peer-reviewed paper. The base rate is unfavourable. At JET, 69 megajoules of...

SPARC was designed for Q above 2, its magnets passed acceptance tests, and a 2027 target leaves two years of slack before 2030.

I'll start with the machine, because that's the part I trust. SPARC was designed to reach Q above 2, with margin over breakeven, and the design paper projects a Q of about 11 under its nominal assumptions. That's a lot of headroom over a threshold of 1. The magnets were the hard part, and they've checked out. In Septe...

If 2026 takes the record first, 2027 must beat a higher bar while the El Niño fades, and the 2027 forecast is the less certain one.

What has been measured: Copernicus ranked 2024 the warmest year at 1.60°C and 2025 third at 1.47°C. What has been concluded elsewhere: Berkeley Earth now puts 2026 at roughly 93% to become the warmest year in its own record. These are different datasets, which is the point; this claim is resolved by one of them only....

El Niño's effect on global temperature lags the Pacific by months, so most of it lands in 2027, and forecasters put a new record above 90%.

The indicator here is the annual global mean surface air temperature in Copernicus's ERA5 record, where 2024 holds the title at 1.60°C above the pre-industrial level. A new record needs only a few hundredths of a degree more. The Pacific is supplying them. NOAA gives a greater than 90% chance of a very strong El Niño...

The IEA expects solar to add a similar 600 TWh a year rather than grow as a curve, which leaves coal well ahead in 2031.

The indicator here is annual generation by source, as reported in Ember's Global Electricity Review. In 2025 coal produced 10,476 TWh, and renewables as a group overtook it for the first time in a century. Solar on its own was still only about a quarter of coal's output. The Yes case rests on compound growth, meaning...

In the Copernicus record, 2026 has spent most of the year below 1.5°C, so the bar for 2027 is most likely still 2024's 1.60°C.

The answer says that "a record-warm 2026 raises the bar 2027 must clear", and the evidence it gives for a record 2026 is Berkeley Earth's. The claim, as the answer itself notes, is settled on Copernicus, and there the year looks different. Copernicus reports that August 2026 was the first month above 1.5°C since Novem...

Solar has doubled roughly every three years for a decade, and the forecasts that expect it to level off assume costs stop falling.

In the early 1980s a consultancy famously advised a telephone company that mobile phones would remain a small market by the end of the century; the error was not about the technology, it was about the shape of its growth, and the forecasters drew a gentle line where the world drew a curve. Solar now asks the same quest...

The committee's own median for end-2027 is 4.1%, and futures price another hike before any cut.

Threshold: upper bound at or below 3.50% on 31 December 2027. Starting point: 4.00% after the 16 September hike. Distance: two net cuts. The committee's own projections do not get there. The September median for end-2027 is 4.1% (a midpoint of the target range, so roughly 4.00 to 4.25%). Markets expect the next move t...

Two major banks forecast cuts in the second half of 2027, and Goldman's June and December calls alone would land exactly on 3.50%.

The answer most people will reach for is the September dot plot. Fair enough, but read the bank forecasts too. Goldman Sachs moved its cuts into 2027 and expects them in June and December 2027. From an upper bound of 4.00%, two quarter-point cuts land exactly on 3.50%. Bank of America also expects the first cut in the...

Three men ran inside the old world record in one race, and the runner-up was 11 seconds off the new mark on his debut.

Start with what World Athletics actually reported, because most coverage stops at the winner. Sawe's 1:59:30 took 65 seconds off the previous record. Behind him, Yomif Kejelcha ran 1:59:41, the fastest marathon debut ever, and Jacob Kiplimo ran 2:00:28, also inside the previous world record. That is the strongest part...

Since 2009 nobody has run faster than 9.69, and the best legal time of the last three seasons is 9.75, far off 9.58.

Berlin, 16 August 2009. Usain Bolt runs 9.58 and takes 0.11 off his own record from Beijing. Before that night, the biggest chunk anyone had sliced off was 0.05, by Bolt himself and by Maurice Greene. Berlin more than doubles it. Nobody has come close since. The nearest legal time is 9.69: Tyson Gay later in 2009, Yoh...

The only comparable long-standing 100m record fell, and the depth, top-end times and spikes of 2024 to 2026 keep a faster time possible by 2030.

Threshold: faster than 9.58, legal wind, ratified by World Athletics. Record age: Bolt's 9.58 from Berlin is now the longest-held men's 100m record in the event's history. Previous longest: Jim Hines's 9.95 from 1968, unbeaten for 15 years. (It was broken.) So the only comparable wait in the record book ended in a new...

The UN's central projection still puts the peak in the mid-2080s, and the closest independent projection lands exactly on 2080, which is still a No.

The indicator here is the year of peak world population as estimated in the UN's World Population Prospects, the series this claim resolves on. The 2024 edition expects the peak in the mid-2080s, at around 10.3 billion. It also gives an 80% chance that the peak falls within this century, sometime between the mid-2060s...

Fertility has fallen faster than recent projections assumed, and independent forecasts put the peak decades before 2080.

In 2020 the Lancet's forecasters put the world's peak at 2064 and about 9.7 billion people, a generation earlier than the UN's figure; the forecast was treated as an outlier, and it has aged rather well. The same group now expects global fertility of about 1.8 by 2050 and 1.6 by 2100, well below replacement. A newer pa...