lead · Thomas Claburn/The Register
TypeSafe AI Introduces Jev for Typed Software Decisions
Confirmed: TypeSafe AI has introduced Jev, a closed-weight model available through early access. The company says Jev accepts structured questions and returns typed decisions with probability distributions and confidence scores for software to use directly.
Why it matters: Typed outputs could reduce the need to parse natural-language responses before software acts on them. TypeSafe's speed, cost and reliability claims have not yet been independently established.
Our read: Jev may be useful for narrow automation tasks where output structure matters, but its early-access status means there is not yet enough public evidence to judge its accuracy or operational reliability.
Open the source for story 1
business · Reuters
Morgan & Morgan Commits $1 Billion to AI and Technology
Confirmed: Morgan & Morgan says it will invest at least $1 billion in AI and technology over the next decade and plans to offer its MX2 platform to other law firms by the end of 2027. Reuters also reports that two of the firm's lawyers were previously sanctioned over AI-generated fake citations; the firm says it has since strengthened training and review processes.
Why it matters: The announcement shows both the scale of legal-sector AI investment and the need for human verification when generated material affects court filings or client work.
Our read: Turning MX2 into a product could create a new revenue stream, but adoption will depend on demonstrated accuracy, security and governance rather than investment size alone.
Open the source for story 2
technology · SemiAnalysis
SemiAnalysis Estimates 2.3 GW Delayed by Data-Centre Moratoriums
Confirmed: SemiAnalysis estimates that local moratoriums and New York's executive order have directly delayed about 2.3 GW of planned US data-centre capacity. The estimate covers policy delays that the firm could tie to individual projects; it does not measure every grid, power, equipment or construction constraint.
Why it matters: The analysis shows why the number of restrictions is a poor proxy for delayed capacity and why project-level evidence is needed to assess policy effects.
Our read: The finding suggests that the direct effect of the mapped moratoriums is narrower than headline counts imply, but it does not establish that the wider US data-centre buildout is free of infrastructure constraints.
Open the source for story 3
Practical takeaway
Treat new-model performance claims as hypotheses until independent testing is available. For operational decisions, verify AI-generated work and distinguish delays caused by moratoriums from wider grid, power, equipment and construction constraints.
What to watch next
Watch for independent Jev evaluations, details of Morgan & Morgan's review controls as MX2 approaches external release, and further project-level evidence on which data-centre constraints actually alter delivery dates.