Missed GTC 2026? You can still catch the best moments. Explore sessions covering physical AI, agentic AI, inference, AI factories, and more now available on NVIDIA On-Demand. 👉 Check out top talks from #NVIDIAGTC 2026: nvda.ws/417WQPM 👉 Access 600+ sessions: nvda.ws/4145PkZ [Translated from EN to English]
GENERATIVE AI
-
Alibaba’s Qwen AI Team Faces Major Leadership Departures
By
–
Did Alibaba just kneecap its powerful Qwen AI team? Key figures depart in wake of latest open source release buff.ly/k3lvow9 #AI #MachineLearning #DeepLearning #LLMs #DataScience
→ View original post on X — @miketamir, 2026-03-30 18:59 UTC
-
Generative BI: Beyond Evolution, a Structural Shift in Data Decision-Making
By
–
Generative BI is not just an evolution of Business Intelligence.
It’s a structural shift in how organizations think, interact, and decide with data.
For years, BI promised democratization. In reality, many companies are still stuck between: IT bottlenecks Low data literacy -
Codex Outperforms Claude Code in Speed and Accuracy
By
–
Damn! Codex finished the task in 2 minutes. Claude Code took 30 minutes and failed.
→ View original post on X — @romainhuet, 2026-03-30 17:22 UTC
-
America Isn’t Ready for What AI Will Do to Jobs
By
–
America Isn't Ready for What #AI Will Do to Jobs
#RiseoftheRobots
theatlantic.com/magazine/202… [Translated from EN to English]→ View original post on X — @mfordfuture, 2026-03-30 16:39 UTC
-
AI Tools Democratization Drives Need for Business Differentiation
By
–
In a world where everyone can build websites, apps and features easily (thank you Cursor, Lovable, Claude and the likes), it will take more for you and your company to differentiate themselves (which is in my opinion the basis for success). That's why we're seeing more and more
-
Data Engineering for Scaling LLM Terminal Capabilities
By
–
On Data Engineering for Scaling LLM Terminal Capabilities buff.ly/AKibXBi #AI #MachineLearning #DeepLearning #LLMs #DataScience
→ View original post on X — @miketamir, 2026-03-30 16:05 UTC
-
Running Operational Workloads with Lakebase, Databricks Apps, Agents
By
–
See how to run operational workloads on the lakehouse using Lakebase, Databricks Apps, and Agent Bricks. This BrickTalks session covers how teams are building data apps and AI agents on top of serverless Postgres to automate workflows and make data usable in real applications.
-
Coding Agents Excel at Processing Massive Long-Context Documents
By
–
// Coding Agents are Effective Long-Context Processors // We are just touching the surface of what's possible with coding agents. LLMs struggle with long contexts, even the ones that support massive context windows. It turns out coding agents already know how to solve this; you just need to reframe the problem. This work places massive text corpora into directory structures and lets off-the-shelf coding agents (Codex, Claude Code) navigate them with terminal commands and Python scripts. This is great, as you are not feeding massive text directly into a model’s context window or relying on semantic retrieval. Results: – On BrowseComp-Plus (750M tokens), this approach scores 88.5% vs 80% best published. – On Oolong-Real (385K tokens), 33.7% vs 24.1%, a 56% relative improvement. – GPT-5 full-context baseline only manages 20% on BrowseComp-Plus. Works up to 3 trillion tokens. Instead of scaling context windows or building retrieval pipelines, coding agents that already know how to navigate file systems can process virtually unlimited context. The agents autonomously develop task-specific strategies: writing scripts, iterative query refinement, and programmatic aggregation. Paper: arxiv.org/abs/2603.20432 Learn to build effective AI agents in our academy: academy.dair.ai/
-
PyTorch trunc_normal_ initialization bugs in LLM training code
By
–
Okay LLM + PyTorch people, trunc_normal_, what the fuck! Many LLM inits use it w/ default cutoffs. It's either not doing anything or it's quite broken due 2 issues. 1. The a/b cutoffs in PyTorch are not in std-devs, they are absolute. So w/ a std=0.02, and -2/2 (default arg) cutoffs that's 100σ!! That is a normal distribution, trun isn't doing anything. 2. There are numerical issues. Even in float32, the truncation produces a handful of -2 (lower cutoff) values, 100σ!! That's incomprehensibly improbable. I doubt a float32 or even float64 algo could even produce it, but clamping a bad float value does. Olmo (@allenai codebases) appear to be one of the few that uses trunc_normal_ and bothered to set the cutoffs properly. It'd be nice to see more train code opened up as a default. We so often only end up with a sanitized version of the inference/fine-tune friendly model these days and may lose details like original init. I've known about #1 for ages, I have an alternate trunc_normal_tf_ implementation in timm for that reason. But I saw those -2's last week when I was debugging something and was a little surprised.
→ View original post on X — @jeremyphoward, 2026-03-30 15:09 UTC