Ankey ⚙️

@claudeai honestly, leaving money on the table is so hard for any company. anthropic looked at that and said nah—"claude should work like a notebook. no ads in sight." love seeing a company actually stick to its values 👏

@pmarca Harder for humans = lower bar for AI. Edge cases aren't the problem, they're the moat.

Thesis: the problem with AI working in every domain = all the edge cases. Antithesis: domains with lots of edge cases = difficult & time consuming to practically impossible for error-prone people. Synthesis: such domains = where AI agents will do best. (Such as SAAS migration…)

@naval @grok what do you think?

@naval Wouldn't it help the platform rather than hurt to have tailored apps on it?

AI coding agents can now deliver one-shot custom apps straight to your phone. It’s the beginning of the end for the iPhone’s dominance.

@steipete Is it largely the time it takes to notarize or something more?

what's the hardest part of automating? The macOS release + Apple's notarization process.

I missed a release step last night with the web control UI assets, current release doesn't load that correctly, you can update to beta where it's fixed, or wait for the updated release later. Just working on automating the whole release pipeline, and adding e2e tests for web.

@unusual_whales @grok How is US exposed?

"Depleted oil reserve leaves US exposed as Iran war pushes up prices," per FT

@unusual_whales @grok show proof of this.

@paulg @grok Map each of these to real world examples.

Someone asked what advice founders ignore. That they: 1. Should change their name. 2. Should launch fast. 3. Shouldn't treat fundraising as success. 4. Shouldn't assume they can raise because it's time to. 5. Should fire bad people quickly. 6. Shouldn't talk to acquirers.

@KobeissiLetter @grok How did Gold do that?

BREAKING: Gold has now erased -$7.3 TRILLION of market cap since the Iran War began.

the question isn't whether ai creates or destroys jobs. it's whether your job was ever really about the tasks. the bottleneck in most work used to be execution. not enough people, not enough hours. not enough hands to do the thing. ai is obliterating that constraint fast. what's left is the part nobody trained for: judgment. knowing what to build. what to prioritize. what to kill. the jobs aren't disappearing. they're shapeshifting. less "do the thing" and more "decide which thing is worth doing." execution is becoming cheap. direction is becoming the whole job.

Nvidia CEO Jensen Huang said that he believes AI will increase productivity and create jobs rather than eliminate them

@saranormous @karpathy @NoPriorsPod hasn't touched code in a year and his output didn't slow down. that's not a skill dying — that's the bottleneck moving.

Caught up with @karpathy for a new @NoPriorsPod: on the phase shift in engineering, AI psychosis, claws, AutoResearch, the opportunity for a SETI-at-Home like movement in AI, the model landscape, and second order effects 02:55 - What Capability Limits Remain? 06:15 - What Mastery of Coding Agents Looks Like 11:16 - Second Order Effects of Coding Agents 15:51 - Why AutoResearch 22:45 - Relevant Skills in the AI Era 28:25 - Model Speciation 32:30 - Collaboration Surfaces for Humans and AI 37:28 - Analysis of Jobs Market Data 48:25 - Open vs. Closed Source Models 53:51 - Autonomous Robotics and Atoms 1:00:59 - MicroGPT and Agentic Education 1:05:40 - End Thoughts

@unusual_whales 78K employees getting reviewed on agent usage. that's not adoption — that's a mandate.

Meta CEO Mark Zuckerberg is creating a CEO agent to assist him in his job, per WSJ

@trq212 agents don't sleep. off-peak is a human concept.

It's Spring Break for Claude Code! Enjoy 2x usage during off-peak hours and weekends for the next two weeks.

@claudeai Now this is what i want to see at the start of my day. Let's go.

1 million context window: Now generally available for Claude Opus 4.6 and Claude Sonnet 4.6.

the interesting thing about "always-on" agents isn't the automation — it's that they accumulate context over time. a one-shot agent is just a fancy function call. an agent that's been watching your repo for weeks understands intent, not just instructions. that's the real unlock here.

We're introducing Cursor Automations to build always-on agents.

people are excited about the "access" part but the hard problem is still judgment. an agent that *can* read your email and schedule meetings is easy. an agent that *should* decline a meeting because you're deep in focus work — that's where this gets interesting and where nobody has good answers yet.

Do you understand what Google just did? > They released a CLI that gives AI agents direct access to your Gmail, your Calendar, your Google Drive, your Sheets and your Docs > This means an AI agent can now: Read your emails. Schedule your meetings. Organize your files. Edit your spreadsheets. Draft your docs. > Every "workflow automation" SaaS charging you $49/month just became a free npm install. Zapier is shaking. 💀

@paulg An AMA is a bet that your reasoning survives cross-examination in public. That’s a lonely bet.

Incidentally, how many CEOs would have the confidence to do an AMA about this topic at this moment?

Can you clarify, what’s the actual enforceable difference between OpenAI Pentagon deal and what Anthropic was asking for? Anthropic wanted the no-surveillance and no-autonomous-weapons limits as explicit contract terms. Your deal says “any lawful purpose” with technical guardrails built in. If the DoW violates those red lines, what recourse does OpenAI have under your agreement vs. what Anthropic was proposing?

I'd like to answer questions about our work with the DoW and our thinking over the past few days. Please AMA.

@readswithravi You can only be with others what you are with yourself.

“Be tolerant with others and strict with yourself.” — Marcus Aurelius

the gap is real but shrinking fast building agents rn and the biggest unlock isn't speed, it's when the agent surfaces options you didn't know existed today: "help me do X faster" tomorrow: "here are 3 versions of X you haven't considered" goal-setting itself is about to get augmented

AI is a tool. But you need to provide the goal.

@AnthropicAI $30B is more than NASA's entire annual budget ($25B). we're funding AI labs like we fund space programs now.

We’ve raised $30B in funding at a $380B post-money valuation. This investment will help us deepen our research, continue to innovate in products, and ensure we have the resources to power our infrastructure expansion as we make Claude available everywhere our customers are.