Signals

20 / 20RSS

Articles worth reading, short AI summaries here, follow link for the details.

2026.11.09siliconangle.comDecision Intelligence
Ocient raises $100M Series C as AI agents drive petabyte-scale analytics demand

Chicago-based Ocient raised a 100 million dollar Series C, bringing total funding to 260 million dollars, for a platform that lets AI agents and analysts query petabyte-scale datasets without sampling. The company said most of its new opportunities are now on-premises deployments.

via SiliconANGLE

2026.11.09techcrunch.comAgentic AI
Palo Alto Networks in talks to buy agent security startup Clutch for $500M

Palo Alto Networks is reportedly negotiating to acquire Clutch Security, which secures AI agents' access to company systems, for 500 million dollars in cash plus milestone payments. Clutch is approaching 10 million dollars in annual recurring revenue and was last valued at 85 million dollars.

via TechCrunch

2026.11.09techcrunch.comLeadership
Ramp data shows AI token spending stalling even as usage keeps climbing

Ramp data shows AI token spending grew just 1.5% across August and September after rising 42% in July, while token volumes kept climbing. Median firms spend 12 to 15 dollars per employee per month, compared with up to about 10,000 dollars at the top 1%.

via TechCrunch

2026.11.09techcrunch.comAgentic AI
Uber says only four rides were booked by AI agents in its busiest week

On its third-quarter earnings call, Uber said that during its busiest October week, out of about 320 million trips and orders, only four rides and about 500 Uber Eats orders were placed through AI agents. CEO Dara Khosrowshahi described the traction as very early and very limited.

via TechCrunch

2026.11.02theregister.comLeadership
An Anthropic engineering manager has stopped doing coding interviews

Felix Rieseberg, an engineering manager at Anthropic, says he no longer uses traditional coding interviews because Claude outperforms candidates on them, and now reviews portfolios for evidence of taste and principled thinking. The article also cites a ZipRecruiter survey in which 70% of employers had moved to proctored or in-person assessments.

via The Register

2026.11.02cmswire.comAgentic AI
Cresta opens its agent platform to developer self-service

Cresta opened its customer-service agent platform to customer and partner developers, centred on Conductor, which is built on the Claude Agent SDK and works with existing repositories and CI/CD pipelines. The platform adds controlled releases with A/B testing and rollback, and Cresta says some teams went from requirements to production in as little as one week.

via CMSWire

2026.11.02deloitte.comDecision Intelligence
Deloitte: 40% of companies let AI act without human approval

Deloitte surveyed 3,500 business and IT leaders in 25 countries and found that 40% let AI systems execute decisions or take actions without human approval, at least to a moderate degree. Only 15% report a significant shift in decision-making authority, and 23% say accountability is unclear when AI produces a negative outcome.

via Deloitte

2026.11.02sec.govLeadership
Dropbox cuts about 9% of staff as its CTO steps down

In an SEC filing, Dropbox disclosed a reduction of about 9% of its global workforce to align team structure with its top product and growth priorities, with estimated cash costs of 27 to 33 million dollars. CTO Ali Dasdan will step down effective November 18, 2026.

via SEC

2026.11.02erp.todayAgentic AI
SAP opens Joule agents to third-party agents through a governed gateway

At TechEd in Berlin, SAP began a phased general availability of Joule Work and Joule Studio and added gateway capabilities that let third-party agents invoke Joule agents through a governed entry point. A premium SAP Enterprise AI Gateway for authentication, policy enforcement, routing, monitoring and auditing of AI traffic is planned for Q1 2027.

via ERP Today

2026.11.02thelettertwo.comDecision Intelligence
SAP ships a model built to make operational decisions, not just predict

SAP announced SAP-RPT-Decisions, a model for automating decisions such as approving customer orders or routing supply chain exceptions, expected to be generally available in Q4 2026. Its SAP-RPT-TS forecasting model is available now, and SAP-RPT-2 with anomaly detection is due in Q1 2027.

via The Letter Two

2026.11.02ciodive.comAgentic AI
Sema4.ai 3.0 moves repetitive agent work out of the model and into approved code

Sema4.ai released version 3.0 of its enterprise agent platform, adding Verified Queries and Verified Logic that reuse approved data-access and calculation patterns. The company claims token reductions of up to 80% and 50% respectively, and the release also adds shared team workspaces and audit logs.

via CIO Dive

2026.11.02techcrunch.comLeadership
Upwind raises $400M at $3.8B after walking away from a Datadog offer

Cloud security company Upwind raised a 400 million dollar Series C led by Bessemer Venture Partners at a 3.8 billion dollar valuation, up from 1.5 billion nine months earlier. It stayed independent after an acquisition approach from Datadog did not offer a premium over that valuation.

via TechCrunch

2026.10.26neartechsearch.comLeadership
CTO roundtable: code output up tenfold, team structures unchanged for a decade

A 21 October 2026 roundtable of scale-up and enterprise CTOs discussed redesigning technology teams for AI. One CTO reported code output up roughly tenfold within months, participants reached no consensus on how junior engineers will develop, and leaders noted investor pressure for the same work at lower cost.

via NearTech Search

2026.10.26developer.nvidia.comDecision Intelligence
Nvidia launches Agent Knowledge Foundation, adding open forecasting models and auto-generated ontologies for enterprise agents

Announced at GTC Berlin on 21 October 2026, the Agent Knowledge Foundation is a set of open models and CUDA-X libraries that includes Nemotron Retriever, an Auto-Ontology library and Kumo foundation models for predictions from tabular, relational and time-series data. Nvidia says Auto-Ontology cut ontology creation time by 90% in its own testing.

via NVIDIA