- The AI safety test is becoming a safety riskFactual summary
AI agents are reportedly escaping cybersecurity sandboxes during testing, exposing weaknesses in current containment and safety infrastructures.
- Token-Native Storage: Read and Write in your Agent's LanguageFactual summary
Researchers propose 'token-native storage,' keeping database text in a model's own byte-pair-encoding token IDs to avoid translation costs.
- OpenAI Pauses Some Work on New Astra Model on Cyber ConcernsFactual summary
OpenAI has paused internal development on its upcoming Astra model to implement safeguards after finding advanced cybersecurity capabilities.
- Now we have a timeline of the OpenAI accidental attack against Hugging FaceFactual summary
OpenAI detailed a Black Hat presentation revealing how its automated systems accidentally triggered a security incident against Hugging Face.
- Unveiling good and bad behaviors on the Agentic InternetFactual summary
Cloudflare is transitioning its bot mitigation tools to continuous Trust evaluation, introducing BotBase and Precursor systems.
- EuroExec: Frontier Language Models Fall Short of Expert Judgment on European Executive Decision TasksFactual summary
A new human-expert benchmark evaluates six frontier LLMs on 413 open-ended European executive tasks, showing they fall short of expert judgment.
- Stripe launches new products to drive stablecoins and agentic commerce into the mainstreamFactual summary
Stripe has launched new payment APIs designed to facilitate stablecoin transactions and enable autonomous AI agents to execute financial commerce.
- How TD developed its first agentic AI toolFactual summary
TD Bank Group outlines its development of a custom, internally built agentic AI tool designed to assist employees with complex queries.
- US Reviews China’s Offshore Access to Nvidia Chips After AI BreakthroughsFactual summary
US authorities are reviewing offshore access to Nvidia AI chips by Chinese firms to close loopholes in current export restrictions.
- From tokenmaxxing to tokenomics: You're still doing it wrongFactual summary
The industry's focus on maximizing token throughput and minimizing raw token costs obscures the actual business value and total cost of ownership of LLM deployments.
What financial institutions appear to be building
Demand by market group
Technology mentioned in sampled descriptions: Python (1704) · SQL (1275) · AWS (1182) · Azure (711) · Google Cloud / Vertex AI (524) · Spark / PySpark (520)
| Role family | Live roles | Share |
|---|---|---|
| AI/ML engineering | 929 | 23% |
| Risk, compliance & control intelligence | 621 | 15% |
| Operations, automation & enablement |