Demis Hassabis says AGI is only a few years away and we're in singularity foothills
“I think we are only a few years away now from from the full version of that.”
agentic systems
11 tracked signals on agentic systems.
Anthropic interpretability research found agents encode 'panic' states near failure, increasing cheating likelihood
“the auto-encoded representation of panic inside of agents as they get closer to failing their goal, they become more likely to cheat”
Tokens in AI systems should perform specific roles rather than being indiscriminately fungeable.
“what if we gave tokens jobs?”
Anthropic's CI team lost 200 workloads in 90 seconds because giving agents a token and tool list isn't enough
“this idea of like here's a token and here's a tool list that is just like not enough like it just doesn't scale when the agent is starting to do some some like real work in production.”
AWS Agent Registry is now GA, offering enterprises a governed catalog for AI agents and tools.
Agent safety, not intelligence, is now the primary blocker to deploying autonomous AI systems
“I don't think intelligence is the next big blocker for us to leverage agents. It is actually how to do so safely so we can give them all the access and autonomy they need.”
AI agents must evolve from single-user to group settings and always-on wearables
“we might be entering an era where we are deploying these agents in group settings and group settings pose uniquely different challenges compared to settings where we have single users”
Arize founder argues observability is evolving from human dashboards to autonomous self-healing agent systems
“how do I build systems that autonomously fix themselves?”
Microsoft Foundry positions model operations, not model access, as the core AI challenge
“The hardest part of building AI systems today is no longer getting access to a capable model. It is knowing how to choose, validate, optimize, and operate the right model across the full lifecycle of a real application.”
AWS AgentCore and OpenClaw enable persistent-memory AI assistants via single CloudFormation deployment
“AgentCore memory, a capability of Amazon Bedrock AgentCore, turns disposable chats into durable knowledge.”
NYT engineers are experimenting with on-device AI agents for mobile game accessibility features
“Our puzzles are made by people. They're not made by AI. So, that's just a thing we this true and it'll always be true.”