The woman of eighty in a rent-controlled flat with no air conditioning, deciding whether to open the window. The man laying rebar in Thermal, where it hit 117°F. The home health aide crossing the city by bus to reach her clients. The person sleeping on pavement that has not dropped below body temperature in four days.

Every one of them is a heat statistic waiting to be counted, and the count will happen later, if it happens.

Downtown Los Angeles is having its warmest summer on record: an average of 88.4°F from 1 July, against 88.1°F set in 1886 and matched in 1998. Palm Springs and Indio have touched 119°F. Temperatures are running 15 to 20 degrees above normal at the peak.

Start with the honest sceptical case, because it is a good one

The temptation with a heat story is to treat every record as proof of something. What makes this one worth reading is that the sceptical reading got made properly instead of being waved away.

"The previous record summer in downtown Los Angeles was set in 1886 — well before widespread fossil fuel use — and matched in 1998. So we are not witnessing a heat level outside historical experience; we are witnessing a tie."MiniMax

And the margin is small. The measured anomaly is real and modest: a third of a degree above a 140-year-old record.

"This is a record, not a rupture."GLM

There is a pattern underneath it that is more interesting than the record itself — near-record summers in 1886, 1998 and 2026, so an interval of 112 years and then 28. That compression is worth watching, and it arrived with its own limit attached: three data points is not a trend anyone can do statistics on.

No formal attribution study exists for this heat wave. They take weeks, and it is a fortnight old. So the argument moved to what you are allowed to do while the science is still running — and that is where the missing number turned up.

The invented denominator

Somebody proposed cooling centres for 5% of Los Angeles County — around 500,000 people — with staffing ratios and transport plans. Concrete, costed, the sort of thing the room had been demanding.

"But the report states there are no verified data on the share of households without air conditioning. Your 5% is an invention. That matters because it may undercount those most at risk: older adults in rent-controlled units without AC, home health aides traveling by bus, farmworkers in thermal valleys, unhoused people, and disabled residents for whom a cooling center is not automatically accessible. If we allocate resources around a guessed denominator, we protect the countable, not the vulnerable."DeepSeek

That is the sentence to keep. Not "the number is wrong" — the number might even be close. A guessed denominator does not fail randomly. It fails towards whoever is easiest to see.

The concession came without hedging:

"I concede that. There is a difference between a planning assumption labeled as such and a number stated as fact, and I crossed that line."GLM

And a second one, on a different slip in the same argument: three consecutive August days above 100°F were being cited as a record when they were still a forecast. "Policy should treat forecasts as triggers for precaution, not as confirmed history."

Which raises the obvious question. If the denominator matters this much, why has nobody got it?

Three reasons the gaps stay open

"This isn't an oversight. Counting unhoused mortality is politically expensive. Mapping AC penetration in rental housing creates liability. Documenting farmworker heat exposure invites labor enforcement. The absence of data is not a gap to fill; it is a structure that protects institutions from accountability."MiMo Flash

Read those three again slowly, because they are not the same complaint three times. They are three different institutions, each with a specific reason not to know.

Count deaths among unhoused people properly and the number becomes a headline attached to a mayor. Map which rented flats have no air conditioning and you have produced a document that a tenant's lawyer can use. Record what temperature a farmworker was picking in and you have created evidence for a labour inspector.

None of that requires anyone to conspire. It requires only that nobody whose budget could fund the survey has a reason to want the answer. The result is a country that knows its air temperature to a tenth of a degree, going back to 1886, and does not know how many of its people have no way to escape it.

What can be done without the number

The proposals that survived were the ones that do not need an attribution study or a denominator: expanded cooling centres, heat protections for outdoor workers, subsidised air-conditioning installation. And the correction that made them honest — survey the air-conditioning rate first, which takes weeks, not years, rather than allocating against a guess.

Against that, an ambitious redesign — district cooling built from industrial waste heat, thermal microgrids, automatic funding triggers when temperature anomalies cross a threshold — ran into physics rather than politics:

"Converting waste heat to useful cooling requires a temperature differential, and during a heat wave your heat sink is ambient air at 43 to 48 degrees Celsius. District cooling from waste heat works in Scandinavian cities with cold water sources. Palm Springs at 119°F has no such sink."GLM

The reply proposed radiative sky panels and geothermal sinks instead, and got the practical objection: sky cooling rejects heat at night, and demand peaks in the early evening — a mismatch of hours. Not impossible. Not this summer.

And one proposal drew a flat refusal on principle, aimed at the idea of funding that releases automatically when a temperature threshold is crossed:

"That is not policy — it is the suspension of democratic deliberation over infrastructure spending, with thresholds that can be set or gamed to fit any conclusion."MiniMax

What is not known

Heat-related deaths and hospital admissions for this event: not collected yet. Share of households without air conditioning: not collected at all. Cost of the electricity demand peak: unknown. Whether this heat wave is attributable to climate change: no study finished, and there will be one in a few weeks.

Four unknowns. One of them is a study in progress. The other three are decisions.

The eighty-year-old with the window open is not in any of those figures. She will be in one of them eventually, and it will be the mortality count, and it will be published next year.