The New Stack
@thenewstack

All about at-scale software development, deployment & management. Tech news, analysis, research, podcasts & more! Subscribe đź“© thenewstack.io/newsletter

 We're Global 🌏
thenewstack.io
Joined February 2014
11,294 Photos and videos
See who spent your AI credits and set fair caps per team via @gitlab about.gitlab.com/blog/new-us…

See who spent your AI credits and set fair caps per team

With GitLab 19.4, set per-user GitLab Credit caps from a settings page, export the session behind a spike, and show each team its own AI consumption.

about.gitlab.com
119
OpenAI makes you call sales for a custom voice. Google just made it self-serve. thenewstack.io/gemini-tts-vo…

OpenAI makes you call sales for a custom voice. Google just made it self-serve.

Google's Gemini 3.8 TTS lets developers design a voice from a text prompt or replicate one from a short clip, then reuse it by ID through the Gemini API.

thenewstack.io
141
Claude Opus 5.5 vs. Opus 5 on reasoning tasks: Cheaper, faster, but not better thenewstack.io/claude-opus-5…

Claude Opus 5.5 vs. Opus 5 on reasoning tasks: Cheaper, faster, but not better

Anthropic says Opus 5.5 is 40% cheaper and 30% faster than Opus 5. On hard reasoning tests, the savings held up. The speed claim fell short.

thenewstack.io
1
265
Your pod may be requesting 25× more CPU than you think — and Kubernetes won’t tell you from @Dynatrace dynatrace.com/news/blog/your…

Your pod may be requesting 25× more CPU than you think — and Kubernetes won’t tell you

You’ve run kubectl describe pod. Your dashboard shows 40 millicores requested for the api-server container. You probably assumed your monitoring had the full picture. The Kubernetes scheduler...

dynatrace.com
240
KubeCon + CloudNativeCon North America 2026: Your Week in Salt Lake City via @CloudNativeFdn cncf.io/blog/2026/09/11/kube…

KubeCon + CloudNativeCon North America 2026: Your Week in Salt Lake City

This November, the cloud native community is heading to Salt Lake City. From November 9–12, 2026, KubeCon + CloudNativeCon North America will bring together adopters and technologists from across...

cncf.io
1
396
WHOOP turned a vulnerability response process that once required an engineering team into an automated workflow built in a single afternoon. See how continuous scanning and automated remediation changed the process. Register to join us live: bit.ly/46lCvZS
1
431
"Impressive level of openness": Xiaomi goes way beyond the usual open-weight playbook with MiMo-V2.6 thenewstack.io/xiaomi-mimo-v…

"Impressive level of openness": Xiaomi goes way beyond the usual open-weight playbook with MiMo-V2.6

Xiaomi streamed its $3.5M reinforcement learning run live and released the weights under MIT. And there's more on the way.

thenewstack.io
1
556
Your AI agent returned a 200, passed its faithfulness check, and still answered the wrong question. The evidence that explains why lives in the trace. Thanks to Dynatrace thenewstack.io/ai-agent-trac…

It passed CI. It passed your evals. The customer still got the wrong answer.

Your AI agent returned a 200, passed its faithfulness check, and still answered the wrong question. The evidence that explains why lives in the trace.

thenewstack.io
501
One bank cut token-processing costs 60% with a Kubernetes stack. But a warning about its resource model raises questions about AI inference economics. Thanks to HPE thenewstack.io/kubernetes-ai…

Kubernetes can run AI inference. But can it count the real cost?

One bank cut token-processing costs 60% with a Kubernetes stack. But a warning about its resource model raises questions about AI inference economics.

thenewstack.io
448
“The first generated token might not be the same as the first visible token,” — Chip Huyen Thanks to @ScyllaDB thenewstack.io/pg-99-conf-20…

Chip Huyen explains how to cut inference costs without new hardware

Ahead of her return to P99 CONF on October 21–22, Chip Huyen’s advice on cutting AI inference costs gets a fresh look in the age of agents.

thenewstack.io
480
Learn how open source helps developers avoid vendor lock-in, protect digital sovereignty, and maintain flexible software architectures. Thanks to SUSE thenewstack.io/avoiding-vend…

Avoiding vendor lock-in through an open-source approach: a developer’s perspective

Learn how open source helps developers avoid vendor lock-in, protect digital sovereignty, and maintain flexible software architectures.

thenewstack.io
579
AI agents don't stop at a locked door. They try the window. That's why agent security has to move from the perimeter to the moment an agent acts. Thanks to Ory thenewstack.io/inside-out-ag…

The agent didn't break your controls. It went around them.

AI agents don't stop at a locked door. They try the window. That's why agent security has to move from the perimeter to the moment an agent acts.

thenewstack.io
2
1
2
615
Anthropic says Opus 5.5 is 40% cheaper and 30% faster than Opus 5. On hard reasoning tests, the savings held up. The speed claim fell short. thenewstack.io/claude-opus-5…

Claude Opus 5.5 vs. Opus 5 on reasoning tasks: Cheaper, faster, but not better

Anthropic says Opus 5.5 is 40% cheaper and 30% faster than Opus 5. On hard reasoning tests, the savings held up. The speed claim fell short.

thenewstack.io
1
746
Azure SRE Agent can investigate incidents and prepare fixes before engineers step in. Getting there takes context, clear permissions, and reliable triggers. Thanks to Microsoft Azure thenewstack.io/azure-sre-age…

Agents operate, humans govern: Scale your operations and reduce toil with Azure SRE Agent 

Azure SRE Agent can investigate incidents and prepare fixes before engineers step in. Getting there takes context, clear permissions, and reliable triggers.

thenewstack.io
472
Claude couldn't hack OpenAI. Then Anthropic shipped Opus 5. thenewstack.io/claude-exploi…

Claude couldn't hack OpenAI. Then Anthropic shipped Opus 5.

Security researchers used Claude Opus 5 to chain two vulnerabilities into a path from a forum image upload to OpenAI's internal GitHub repository in under 72 hours.

thenewstack.io
810
A powerful reranker can't rank what retrieval never finds. Our upcoming live session explores why improving the first stage of AI search can matter more than simply adding a bigger model downstream. Register here: bit.ly/3UMqBG4
502
Q.ANT gives away the software for its light-powered AI chips in a CUDA-style bet on developers thenewstack.io/q-ant-open-so…

Q.ANT gives away the software for its light-powered AI chips in a CUDA-style bet on developers

Developers can try it on a regular laptop today. Access to the actual chips opens up over the coming months.

thenewstack.io
593
Scaling AI agents means solving problems that aren't always visible during a pilot. Join Ravi Marwaha, COO and Chief Product & Technology Officer at @arangoai, and TNS Host Dr. Kim Fessel for a look at the six requirements every production-ready agent architecture needs. Join us: bit.ly/3UWIB0d
477
OpenTelemetry everywhere: Migrating a metrics platform at scale via @CloudNativeFdn cncf.io/blog/2026/09/17/open…

OpenTelemetry everywhere: Migrating a metrics platform at scale

For most of the last decade our metrics pipeline ran on gostatsd, the open-source StatsD implementation we maintain. It primarily did two jobs: as sidecar on every host and the aggregation tier at...

cncf.io
2
1
1,043
Amazon blocked Meta's Muse. Then Shopify wired it into every store. thenewstack.io/amazon-meta-m…

Amazon blocked Meta's Muse. Then Shopify wired it into every store.

Amazon kicked Meta's Muse off its store because the agent never said what it was and seemed to be holding onto people's passwords. A day later, Shopify let Muse into every one of its stores. Why the...

thenewstack.io
836
Load more