Every founder I talk to is running the same experiment right now: Pour money into AI, watch the team move faster, then squint at the numbers trying to work out whether any of it is actually paying off. Samar Abbas, Temporal's CEO, is just more honest about the squinting than most.
Temporal sells the reliability layer under a lot of the AI you already use, and this year Abbas turned it on his own company.
AI spend shot up 5x, and revenue doubled. And by his own account, there’s no proof the two are connected. I called him to find out how a company whose whole brand is "reliable as gravity" learns to move fast without drowning in AI slop, and what he still can't measure.
The conversation ran from coding agents and cost controls to AI slop and the reliability layer he's betting the agent era will run on. But it kept circling the one number he doesn't have: proof that any of it reaches the customer. For a company that sells certainty, that's a lot of faith.
What comes after attention? This startup says it already knows.
Subquadratic built a model that claims 1,000x less compute at long context. Its CTO, Alex Whedon, says that's not even the interesting part! He spoke to The New Stack about the "zero attention" architecture he thinks will replace it, and why he expects it to make his own company's breakout model obsolete.
Your agents are hallucinating. Here's how to fix it.
Logs, metrics, and traces work great until an AI agent starts hallucinating or overconsuming. In this session, Dotan Horovits (CNCF Ambassador) and Rekha Thottan (AWS) walk through how to monitor traditional services and evaluate AI agent behavior in production, using a single open-source stack built on OpenSearch and OpenTelemetry.
Chas Clawson ran Red Team assessments at the NSA before architecting the SIEM stack for a federal security operations center. In this insightful webinar, he’s telling us what that taught him about AI, alerts, and where most SecOps teams still get it wrong.
Now in its 9th year, the AI Infra Summit is the premier event for engineers and AI/ML practitioners building full-stack AI infrastructure. Join 8,000+ attendees for cutting-edge hardware and data center innovations, early access to announcements from AWS, Google, Meta, and Microsoft, and direct access to the peers and investors shaping AI infrastructure's future. TNS readers get 15% off with code THENEWSTACK15.
Your access strategy wasn't built for AI agents calling APIs and querying databases across your infrastructure. Explore what a unified access model actually looks like in production. Here's what you'll learn:
One policy for humans, pipelines, and AI agents alike.
Tailscale's stack in action: Border0+Tailscale and Aperture.
.png?height=72&upscale=true&name=The%20New%20Stack%20logo%20-%20Reversed%20(2).png){kind=small}
