One operator's 30-day log counts 6,027 requests from Meta's training collector, 155 from the crawler documented for Meta AI search quality, and zero from the fetcher that fires when a person asks a question.
Reality
- Evidence34
- Adoption20
- Hype gap+8
- Incentives62
- Confidence44
UNICEF's chief statistician says six frontier models averaged 21.2% accuracy across more than 133,000 questions about development indicators, and the UN's answer is a portal that hands agents the numbers with their sources attached.
Reality
- Evidence52
- Adoption57
- Hype gap+30
- Incentives72
- Confidence58
Nicholas Thompson took The Atlantic from red ink to profit and 1.6 million subscribers. He says the early OpenAI deal has paid in cash and in enforceable limits on imitation, and that the traffic he expected did not follow.
Reality
- Evidence40
- Adoption38
- Hype gap+15
- Incentives72
- Confidence55
A two-person startup claims more than 1,000 stores under agent custody. The disclosure that would test it is missing, and the limit on all of it is a product page machines cannot fully read.
Reality
- Evidence41
- Adoption34
- Hype gap+38
- Incentives74
- Confidence36
The operator of IELTS Writing Checker went from 200 to 1,000 generated pages a day. Google's approximate index count fell from about 3,500 to 2,500 and Search Console impressions hit zero.
Reality
- Evidence28
- Adoption12
- Hype gap+22
- Incentives55
- Confidence30
Conductor's benchmark across 13,770 domains puts AI referral traffic at 1.08% of visits, growing about 1% a month. The measurement problem is larger than the channel.
Reality
- Evidence46
- Adoption28
- Hype gap+12
- Incentives74
- Confidence42