
Peter Steinberger's artificial intelligence agent tool OpenClaw has taken the tech world by storm with its ability to execute real-life tasks such as checking him in for his flight to Tokyo. AI is not yet a ubiquitous personal assistant for ordinary people, but "you'll see much more of that this…
Corey Quinn with the smackdown on AWS announcements again:
> Seven design principles for letting AI loose in financial services. I'd have gone with one: "Don't." But sure, comprehensive observability and fine-grained access controls sound great until an autonomous agent decides to rebalance a portfolio at 3 AM and your only explainability is a CloudWatch log nobody reads.
Nachdem ich meine Websites und meine Programmierfähigkeiten vor mehr als acht Jahren für andere Hobbys zurückgestellt habe, kam mir heute die Idee, mal wieder ein paar kleine Seiten zu erzeugen. Und, was soll ich sagen … mir explodiert gerade fast der Kopf. 🤯 WAS. FÜR. EIN. UNTERSCHIED! Ich fühle mich, als wäre ich vor acht Jahren aus einem Käfer ausgestiegen und heute in einem Porsche Taycan auf die Rennstrecke zurückgekehrt. Was für eine Weiterentwicklung! Geil! 🤩
A month with #OpenClaw has me convinced that Agentic AI is the real deal.
I've been skeptical of the AI hype cycles. Even here, the economics are largely not there yet, and it’s far from consumer-ready. But it's hard to deny the raw potential it is already showcasing.
For a few cents, it can draft a script, validate it, and instruct a user how to run it on a system cron. From there, an agentic cron can retrieve the output and make decisions entirely locally on modest consumer hardware. Its easy to see how chaining this together can build valuable automation.
The question is whether frontier models can become economically (and environmentally) viable enough so that the user doesn't have to think about the architecture. Can we hand tasks off to more reliable models without breaking the bank or the ozone layer? Can we trust the inputs will be private and secure?
#OpenClaw #AgenticAI #LLMs #CyberSecurity #CanTech #SelfHosted
Alert triage is a reasonable place to start with AI. It is not a reasonable place to stop.
Most of what slows down security service providers has nothing to do with triage.
It's the onboarding that takes multiple days. The cross-tenant configuration work nobody wants to touch. The detection rules that need to be written, tested, and deployed across fifty environments.
Vendor AI addresses one use case at a time, which means a separate product, a separate workflow, and a separate wait for the next release.
ASW takes a different approach.
Give it hundreds of tools, describe the outcome you want, and it finds a way to get there. That's not a feature. That's a fundamentally different way of getting value from AI, across the whole operation, not just the alert queue.
Watch the full keynote: https://www.youtube.com/watch?v=QS0DzO2rNJw