Closing a startup now includes deciding whether to sell its Slack history to an AI lab
SimpleClosure says 47% of its startup wind-down deals now include a sale of code or workspace data. One founder netted about $25,000 for his code base, and a lawyer who advises AI startups calls such data "legally radioactive."
Reality
- Evidence50
- Adoption45
- Hype gap+20
- Incentives70
- Confidence50
buildOne report1 publisher Self-hosting Mistral Large 4 means loading 1.05 trillion parameters to run 52 billion per token
Mistral AI put Large 4, a 1.05-trillion-parameter mixture-of-experts model, into public preview on Oct. 6 and promised weights by October's end. Self-hosters can size memory now, though Mistral says the training run behind the preview is still going.
Reality
- Evidence50
- Adoption
- Not scored
- Hype gap+20
- Incentives65
- Confidence55
Google's Gemini 4 Argon tied GPT-6 Astra and Claude Fable 5.1 at 53 on the Artificial Analysis index, though some Google staff say its coding lags. Google disputes them, and until paid API access has a date, buyers cannot check the score on their own code.
Perspective Coverage
6 publishers
- Builder
- Builder 37%
- Operator
- Operator 35%
- Investor
- Investor 28%
Reality
- Evidence55
- Adoption15
- Hype gap+25
- Incentives65
- Confidence55
ZeroDrift's own FINRA benchmark puts its 9B model level with GPT-6 Astra
The startup says Anchor 3.0 caught more than 90 percent of violations in a test it published itself, at under a 500th of the cost of a frontier call, and the messages it misses stay the firm's problem.
Reality
- Evidence30
- Adoption16
- Hype gap+32
- Incentives82
- Confidence58
A Shenzhen architect's second shift teaches AI to write her building proposals
Falling infrastructure work halved her income, so she logs into TalentsAI after hours and documents her own reasoning for a model that may one day do the job. Tasks pay 100 to 500 yuan, when they pass review.
Reality
- Evidence55
- Adoption58
- Hype gap+10
- Incentives68
- Confidence55
China's top six AI labs spend about $500m a year on data labelling from US firms
The money is small measured against these vendors' own revenue and large measured against anything else in the training-data supply chain, which is why a control that does not yet exist would be cheap to impose and hard to enforce.
Reality
- Evidence20
- Adoption30
- Hype gap+45
- Incentives
- Not scored
- Confidence22