Live from ISC West 2026. NVIDIA and our partners are showcasing how vision-language models (VLMs) and AI agents reimagine operations by turning video streams into real-time insights. 🧵 1/5 [Translated from EN to English]
ENTERPRISE AI
-
Architecture determines predictive maintenance insight timing
By
–
The architecture determines whether predictive maintenance insights arrive in time to prevent failures or after the damage is done.
-
Real-time telemetry platforms for operational intelligence systems
By
–
Purpose-built platforms handle high-velocity telemetry at the resolution that real-time operations actually demand.
-
Time Series Databases: Architecture and Deployment Strategies
By
–
Time series databases index data by time as the primary axis.
— Lucian Fogoros (@fogoros) 26 mars 2026
Every layer differs: ingestion speed, compression efficiency, query latency, cardinality support. Edge, cloud, or hybrid deployment.
Partner content with @InfluxDB. #influxdata_iiot #InfluxDB pic.twitter.com/tFCXY19IEdTime series databases index data by time as the primary axis.
Every layer differs: ingestion speed, compression efficiency, query latency, cardinality support. Edge, cloud, or hybrid deployment.
Partner content with @InfluxDB
. #influxdata_iiot #InfluxDB -
Factory 2030: Software-Defined Transformation Over Hardware Cycles
By
–
The factory floor of 2030 will be defined by software, not hardware generations. The companies building toward that future are not waiting for the next full hardware replacement cycle. They are starting the software-defined transition now, with the assets they already have.… pic.twitter.com/cgyWfF9112
— Lucian Fogoros (@fogoros) 26 mars 2026The factory floor of 2030 will be defined by software, not hardware generations. The companies building toward that future are not waiting for the next full hardware replacement cycle. They are starting the software-defined transition now, with the assets they already have.
-
Evals Become the Moat as Teams Build on Frontier Models
By
–
An interesting implication of your second category: evals are becoming the moat. If everyone ships on the same frontier model, the teams that figure out how to encode their taste and requirements into automated scoring will compound quality gains over time. Teams that skip this
-
LangSmith Fleet Integrates Microsoft 365 Tools
By
–
LangSmith Fleet now supports @Microsoft365 tools.
— LangChain (@LangChain) 26 mars 2026
Now you can:
→ Build an agent that triages your Outlook inbox and drafts replies
→ Access Sharepoint docs so your agent has the context it needs
→ Connect any agent to Teams so your whole team can use it
Try Fleet:… pic.twitter.com/ObmwvlhXvTLangSmith Fleet now supports Microsoft 365 tools. Now you can:
→ Build an agent that triages your Outlook inbox and drafts replies
→ Access Sharepoint docs so your agent has the context it needs
→ Connect any agent to Teams so your whole team can use it Try Fleet: -
Engineers Hiring Now: Roles, Locations, and Applications
By
–
Who’s hiring engineers right now? Reply with the role, location, and how to apply.
-
AI in RAN: Three Distinctions and Measurable Performance Gains
By
–
First, let’s clarify something most people miss: There isn’t just “AI in telecom.” There’s: → AI in RAN
→ AI for RAN
→ AI on top of RAN AI in RAN is where things get tangible. We’re talking: → 20% improvement in uplink spectral efficiency
→ 14% energy savings with -
Telecom Network Autonomy: Moving Beyond Real-Time Data Limitations
By
–
Most telecom operators are optimizing their networks with 15-minute-old data.
— Ronald van Loon (@Ronald_vanLoon) 26 mars 2026
Offline files.
Batch transfers.
Human review.
And we wonder why autonomy is stuck at Level 3.
At MWC, I sat down with Joe Constantine from @ericsson to unpack what it really takes to move from… pic.twitter.com/kh4X2dp7etMost telecom operators are optimizing their networks with 15-minute-old data. Offline files.
Batch transfers.
Human review. And we wonder why autonomy is stuck at Level 3. At MWC, I sat down with Joe Constantine from @ericsson to unpack what it really takes to move from
