Joe Biden is a leader who acts with the best interests of America in mind. Not seeking re-election is one of the most selfless acts we’ve seen from a politician in modern American history. This is what’s right for our country—and our democratic future.
Kamala Harris is the right person at the right time. Donald Trump and JD Vance are promising an agenda that will wreak havoc on the American people. Harris’s background and leadership growing the economy, fighting for bodily autonomy, and protecting our democracy uniquely position her to push back against Trump’s extremism.
The Democrats are the party of policy, progress, and action. When presented with the choice between Kamala Harris and Donald Trump, I believe in the American people to make the right decision for our country. The Biden-Harris administration has put this country on the right track. It’s time for us to unite. I wholeheartedly support Kamala Harris and her candidacy for President of the United States in our fight for democracy in November.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
Most people are focusing on Chief AI Officers and pilot programs, when the real value is in the unglamorous work where organizations bleed time.
More thoughts:
Start with the coordination layer. It’s the highest-leverage, lowest-drama place to deploy AI.
The biggest language workload inside any enterprise is the coordination layer: Meetings, notes, docs, action items, status updates, etc.
The goal is turning the organization’s memory into something structured and retrievable, so you stop relying on whoever happened to be in the room to (1) achieve the objective discussed and (2) decide on who else should be informed.
Trump should release all of the Epstein files: every person and every document in the files.
I want this complete release because it will bring justice for the victims.
I want this complete release because it will show that the calls for baseless investigations of me are nothing more than political persecution and slander.
I was never a client of Epstein’s and never had any engagement with him other than fundraising for MIT.
1/ I want to state plainly: in all industries, especially in AI, it’s important to back the good guys.
Anthropic is one of the good guys.
More thoughts about why we need to fuel innovation and talk safety at the same time:
I usually try to avoid commenting directly on the hyperscalers (I led OpenAI's first round, and Greylock is an investor in Anthropic).
But I feel very strongly that it is important to comment here. I would say the below as strongly even if no investment existed.
Anthropic, along with some others (incl Microsoft, Google, and OpenAI) are trying to deploy AI the right way, thoughtfully, safely, and enormously beneficial for society.
That’s why I am intensely rooting for their success.
1/ A recent Stanford study led by @erikbryn found that entry-level jobs for 22-25 year-olds in fields most exposed to AI have dropped 16%.
Some reactions to the data, and why I believe we need to design a new on-ramp to work in the AI era:
2/ Any job where you ask a human to act like a robot, a robot will eventually do the job better. Customer service roles (scripted, repetitive) are textbook cases.
And so it’s not surprising to me that these jobs are where we’re seeing some replacement with AI.
3/ The more interesting puzzle is the drop in junior engineering roles. My belief is that all human information work will have a software copilot.
ChatGPT may be the first AI that most of the 8 billion people on our planet use.
Some people don't understand why OpenAI took the risk of deprecating all previous models in favor of GPT-5.
It’s a full-throttle, no-looking-back blitzscale bet. And here’s why it may win:
1/ By opening GPT-5 to everyone immediately, OpenAI locks in massive network effects. New users flock in, existing users upgrade en masse, and ChatGPT’s market pull intensifies. Yes, they’ll spend more to service the use of their most powerful model.
2/ GPT‑5 is priced at $1.25 per million input tokens and $10 per million output tokens—about half the input cost of GPT‑4o. The upside, though, is that they’ll continue to grow the number of new users integrating ChatGPT into their daily lives.
It's not enough to have a firm grasp of the industry you're building in.
If your product doesn’t speak to core human impulses, it probably won’t scale.
Here's why builders need to have a strong theory of human nature:
Founders start with a good idea or a novel technology, but many haven't articulated how that idea aligns with what it means to be human.
If you're trying to create a mass-market consumer application, you have to ask yourself: "What are the parts of what it is to be human that I am triggering, responding to, surveying, etc?"