Rebecca Mendanhall

Columnist @ Bloomberg | Prev: @Nasdaq | Mail: [email protected] | Weekend explorer 🏔️

[ClaudeDevs @ClaudeDevs]

Claude Code on desktop now has an in-app browser.

Claude can pull up docs, designs, or any other site. It can read, click through, and interact the same way it does with your local dev servers.

It's sandboxed and configurable: you choose whether sessions persist.

[ClaudeDevs @ClaudeDevs]

We've reset 5-hour and weekly rate limits for all users.

[Claude @claudeai]

We're extending access to Claude Fable 5 on all paid plans through July 12.

[Anthropic @AnthropicAI]

New Anthropic research: A global workspace in language models.

Of everything happening in your brain right now, only a tiny fraction is consciously accessible—thoughts you can describe, hold in mind, and reason with.

We found a strikingly similar divide inside Claude.

[Claude @claudeai]

Introducing a new way to reflect on how you use Claude.

Your monthly recap shows when you use Claude most and what you spent that time working on, with options to set quiet hours and nudges to take breaks. Find your dashboard in Settings under Reflect: https://t.co/8QAn47W5rI

[Thariq @trq212]

https://t.co/hPiZr1kG7r

[Claude @claudeai]

There’s hope in hard questions.

[Jock Ferguson @jock_ferguson]

Meet Claude for Marketing.

Enter your website and Fastlane will create thousands of videos and social media accounts promoting your product in one click.

This is actually insane.

[Boris Cherny @bcherny]

New in Claude Code: /checkup

Run /checkup to:

1. Clean up unused skills/MCPs/plugins and save context
2. Dedup your local CLAUDE.md against the checked in CLAUDE.md
3. Break up root CLAUDE.md into nested CLAUDE.md's + skills
4. Turn off slow hooks
5. Update your Claude Code to the latest version
6. Enable auto mode by default
7. Pre-approve frequently denied read-only commands

.. And a few other goodies.

/checkup confirms with you before making any changes. Enjoy!

[Rebecca Mendanhall @ChristianRyanml]

@davidmarcus @lightspark i would love to cover this. are you open to press coverage?

[Jeremy Allaire - jerallaire.arc @jerallaire]

The Task Economy is the Agentic Economy

Agents as a service

Consumption pricing

Agentic economic contracts

[Tom McGrath @banburismus_]

we've found a way to discover neural geometry, completely unsupervised, and the results are incredible! this is a really major step forward in interp in my opinion. there's so much structure in there

[Felix Rieseberg @felixrieseberg]

Big day! Claude Cowork is coming to web and mobile, so Claude can keep working while your computer is closed.

This is a major update to Cowork. It combines the power of giving Claude access to your context, an advanced loop for long-running tasks, and the convenience of not needing your laptop to be open.

[Yoonho Lee @yoonholeee]

Thanks for discussing meta-harness @lilianweng !

A very clearly written summary of the state of harness engineering + RSI. I often get the question"what can a harness actually improve?" which i only had partial answers to; this post lays out the whole picture in one place

[Greg Brockman @gdb]

Very excited to help chart the future of Git (and SCM generally) for the agentic future with Taylor!

[Rebecca Mendanhall @ChristianRyanml]

@kernio Are you guys open to the press by any chance? I'd be happy to cover on this.

[Everett Randle @EverettRandle]

https://t.co/KLXJRNdpOl

[Vincent Hus @vincent_hus]

Production AI is broken in ways nobody has fixed yet.
We're fixing it in the open. Apache 2.0.

[Praveen Neppalli @praveenTweets]

Agentic AI adoption is on fire at @Uber, and it's changing the way we build, not just in engineering, but across the entire company.

Today, 99% of our engineers use AI tools. More than 70% of pull requests are attributed to local or cloud agents. And our engineers have built 2,500+ agent skills across the software development lifecycle.
Those numbers are exciting, but they led us to a much bigger question:

How do we bring agentic AI beyond engineering?

Finance. Legal. Operations. Marketing. Customer Support. HR. Procurement.

These functions run on complex workflows that are often manual, highly nuanced, and spread across dozens of systems. You can't automate them effectively by looking at process diagrams or documentation. You have to understand how the work actually gets done.

So we created something called Agentic Pods.

The idea is simple.

We handpicked ~30 of our most AI-proficient engineers (people with deep knowledge of Uber's systems) and paired each of them with a domain expert from a business function.

Then we gave every pod just two weeks.
• Days 1 – 2: Shadow the expert. Observe every step. Document workflows. Ask questions. Build intuition.
• Day 3: Prioritize opportunities based on scale, repetition, business impact, and data availability.
• Days 4 – 5: Build a working agent alongside the person doing the job.
• Days 6 – 9: Validate with several others performing the same work. Does it generalize? Does it actually make their job better?
• Day 10: Ship.

In just the past two months, we've run 16 Agentic Pods across 16 different business functions.
• Capital allocation across 150 cities: 15 hours → 30 minutes.
• Financial pacing reports: 2 days → 10 minutes.
• Marketing web quality assurance: 2 weeks → 50 minutes.
• Support workflow creation: 9,000 manual workflows → self-service automation.

The productivity gains are impressive, but what surprised us most wasn't the speed.
• It was how quickly engineers embedded in unfamiliar domains uncovered opportunities that had been hiding in plain sight.
• The biggest wins rarely come from automating one task. They come from rethinking an entire workflow. Once you redesign the workflow around AI, you often eliminate handoffs, remove unnecessary approvals, replace legacy tooling, reduce vendor spend, and dramatically accelerate decision-making.
• The workflow becomes the unit of automation - not the individual task.
• The most impactful agent skills cut across teams, orgs, functions, tools, and systems.

The biggest lesson? The best AI opportunities are rarely visible from the outside.

You discover them by sitting next to the people doing the work, understanding every friction point, and building with them, not for them.

We're now forming a dedicated team to scale this further and go deeper. They'll deeply understand the work, redesign it from the ground up, and use AI to fundamentally change how the business operates.

It's exciting times!

[Brett Harrison @BrettHarrison]

Compute options will be as transformative to the US AI industry as compute futures. Sellers and buyers of compute won't only use options for insurance, they'll write options and get paid. Covered calls and cash-secured puts:

Compute sellers: covered calls

Compute sellers such as neoclouds are long inventory: racks of accelerators that will be rented out over the coming years. Against the capacity they already own, they can write covered calls on GPU-hour prices above today's market and collect the premium as income. If prices stay flat or drift lower, they keep the premium as yield on top of their rental revenue. If prices rise above the strike price at expiration, they rent out capacity at the strike price while still keeping the premium. The upside beyond the strike is capped, but in return the seller converts a portion of future GPU price volatility into present cash. At each expiry, the seller can keep writing calls against inventory, harvesting premium from a depreciating asset.

Compute buyers: cash-secured puts

Compute buyers such as frontier model companies and AI labs are short inventory, looking to acquire rental time on GPUs. Buying futures locks in their effective price, and buying calls insures against rising prices. But there's another important derivatives trading strategy for compute buyers: writing cash-secured puts. If there's a price below the prevailing GPU rental rate that a buyer would be happy to pay, they can sell a put at that strike and immediately collect the premium as income. If prices rise, they keep the premium when the put expires out of the money. If prices fall to the strike or below, they effectively buy compute at the lower price, partially subsidized by the premium they collected.

Both strategies, executed with discipline and proper risk management, generate income while the worst-case exercise scenarios align with the natural hedge.

In summary:
• Futures: lock in the forward price of compute.
• Buying options: insure against adverse moves in GPU prices.
• Selling covered calls or cash-secured puts: generate income where exercise aligns with the natural hedging outcome.

These strategies are hallmarks of mature US derivative markets facilitated by CFTC oversight. The American Innovation Exchange and our industry partners are ready to support compute as a US-developed asset class from the beginning.