← 2026-03-03

Daily Edition

2026-03-04

2026-03-05 →

X / Twitter

11
Aaron Levie
Aaron Levie @levie
At Box, we just reported our Q4 and full FY26 results, coming in at $1.18B in revenue for the year, and we guided to $1.275B for FY27.

This growth continues to be driven by the need to accelerate knowledge work with the power of AI agents in the enterprise, and the role that content plays in an agentic strategy. Thanks to our customers, partners, and Boxers for driving these strong results.

As I just shared on our earnings call, nearly every enterprise leader that I talk to is looking to transform how their company operates with AI agents. As they go down this journey, they quickly find that for AI agents to be effective in a workflow, they need critical context about the business: the company's product roadmap, marketing strategy, sales materials, financial models, HR policies, best practices, strategy decisions, and so on.

Much of that unique context lives inside of enterprise content, ranging from contracts and financial documents to research documents and marketing assets, housed inside of PDFs, docs, presentations, media files, spreadsheets, and markdown files. All of this enterprise content is the digital brain of an organization.

And as we prepare for a world where there will be 100X more agents inside of an enterprise than people, we can expect to see incredible growth in this unstructured data. Files are, quite simply, the native unit of work for agents. Agents use files to keep track of their work, they leverage files as context about the tasks they’re doing, and use them to share back and forth with their human counterparts.

And as Agents help us augment all of our work across industries like pharma, financial services, legal, or healthcare these agents will need the same level of security, data governance, auditability, logging, and access controls that people need.

Thus, to have an effective AI agent strategy, companies fundamentally need a content strategy. We're building the platform to solve this at Box, and our focus is powering how content gets securely connected to people, agents, and apps.

To support these growing AI use-cases, we’re making it as easy and secure as ever to leverage Box as a file system for AI, connecting content to the entire AI stack in a company -- including Claude Cowork, OpenClaw, ChatGPT and Frontier, Perplexity Computer, Cursor, Copilot, IBM watsonx, Agentforce, or any custom custom agent built with Claude Agent SDK, Langchain, and many more-- via Box’s APIs, MCP server, and CLI support.

We're still insanely early in what this all looks like, but we're incredibly excited to be aggressively building new capabilities for customers and developers to fully take advantage of content + agents to accelerate knowledge work. Excited about the path ahead!
Kevin Weil 🇺🇸
Kevin Weil 🇺🇸 @kevinweil
Retweeted
Mehtaab Sawhney Mehtaab Sawhney
Last week in a beautiful preprint, Dmitrii Zakharov proved a bound for a two-family version of a word-overlap problem, but the first version left a constant factor gap of e. That gap was quickly improved by a mix of GPT-5.2 Pro and human collaborators. Even more recently, an internal OpenAI model obtained the asymptotically sharp bound. Dmitrii incorporated the argument into v2 of his paper https://arxiv.org/pdf/2602.20143; see his AI statement for more details.
Ryo Lu
Ryo Lu @ryolu_
Use Cursor anywhere!

Cursor: Cursor is now available in JetBrains IDEs through the Agent Client Protocol.

http://cursor.com/blog/jetbrains-acp
Aaron Levie
Aaron Levie @levie
Retweeted
Box Box
In our latest partner podcast episode, @BenAtBox, CTO of Box, sat down with @ankrgyl, CEO of @braintrust, to explore how organizations can effectively evaluate, test, and deploy AI agents at scale.
Timestamps
00:39 Ankur Goyal shares his journey from AI document processing to Braintrust
03:01 Defining evals and how they work in AI
07:03 Non-determinism and complexity in AI agents' decision-making
15:12 Advice on handling non-determinism when working with financial data in AI
17:40 Using multiple paths for validation and the importance of cross-checking results
22:12 The critical role of context in evaluating AI output accuracy
26:03 Internal evals as the cornerstone of reliable AI product development
32:16 Promoting transparency in AI evaluation with vendors
34:45 Advice for enterprises to avoid failure when deploying agentic capabilities
Kevin Weil 🇺🇸
Kevin Weil 🇺🇸 @kevinweil
💥 AI accelerating high energy physics

Just a few weeks after the gluon scattering paper, this morning we posted the more complicated graviton scattering analogue. See below for more from @ALupsasca 👇

Alex Lupsasca: We just posted a new preprint: “Single-minus graviton tree amplitudes are nonzero.”

Yes: a helicity sector long assumed to vanish in quantum gravity can actually appear under well-defined kinematics. Preprint: https://cdn.openai.com/graviton/graviton/graviton.pdf
Andrew Ng
Andrew Ng @AndrewYNg
New course: Build and Train an LLM with JAX, built in partnership with @Google and taught by @chrisachard.

JAX is the open-source library behind Google's Gemini, Veo, and other advanced models. This short course teaches you to build and train a 20-million parameter language model from scratch using JAX and its ecosystem of tools.

You'll implement a complete MiniGPT-style architecture from scratch, train it, and chat with your finished model through a graphical interface.

Skills you'll gain:
- Learn JAX's core primitives: automatic differentiation, JIT compilation, and vectorized execution
- Build a MiniGPT-style LLM using Flax/NNX, implementing embedding and transformer blocks
- Load a pretrained MiniGPT model and run inference through a chat interface

Come learn this important software layer for building LLMs!

https://www.deeplearning.ai/short-courses/build-and-train-an-llm-with-jax/
AndrewYNg
AndrewYNg @AndrewYNg
New course: Build and Train an LLM with JAX, built in partnership with @Google and taught by @chrisachard.

JAX is the open-source library behind Google's Gemini, Veo, and other advanced models. This short course teaches you to build and train a 20-million parameter language model from scratch using JAX and its ecosystem of tools.

You'll implement a complete MiniGPT-style architecture from scratch, train it, and chat with your finished model through a graphical interface.

Skills you'll gain:
- Learn JAX's core primitives: automatic differentiation, JIT compilation, and vectorized execution
- Build a MiniGPT-style LLM using Flax/NNX, implementing embedding and transformer blocks
- Load a pretrained MiniGPT model and run inference through a chat interface

Come learn this important software layer for building LLMs!

https://www.deeplearning.ai/short-courses/build-and-train-an-llm-with-jax/
Andrew Ng
Andrew Ng @AndrewYNg
Apple just named its latest laptop Neo -- same name as my son! Should I buy one?

If I run Amazon Nova on an Apple Neo I hope to blow both of my kids' minds.
AndrewYNg
AndrewYNg @AndrewYNg
Apple just named its latest laptop Neo -- same name as my son! Should I buy one?

If I run Amazon Nova on an Apple Neo I hope to blow both of my kids' minds.
Kevin Weil 🇺🇸
Kevin Weil 🇺🇸 @kevinweil
We integrated the Codex harness into Prism (http://prism.openai.com) — this means you get skills, reasoning levels, and the raw tenacity of the Codex model in your LaTeX environment.

Oh and we also built version mgmt into Prism, which was one of the top requests.

See below thread from @vicapow for some great examples of the power of Codex inside Prism 👇

Victor Powell: 🧵1/ We've brought the most advanced AI to Prism by introducing Codex to Prism. Prism is already the best place for scientific writing to happen—and with Codex, now you can write, compute, analyze, and iterate all in one place.

Ryo Lu
Ryo Lu @ryolu_
Notion + Cursor = ☯️

Geoffrey Litt: ✨New demo: what if vibe coding felt more visual?

@brian_lovin @maryrosecook and I did a game jam using Notion as our "IDE": launching Cursor agents from a task board, and making a custom image for each task 😎

The demo shows 3 ideas for the future of agents:

1) Agents should

YouTube

0

No recent videos fetched on this date.