To understand what happens next with AI.... Imagine the world before electricity. For centuries, every home, factory, and office that needed mechanical power had to have its own dedicated “power plant” — a water wheel, windmill, steam engine, teams of horses, or human laborers turning cranks. These were huge, expensive, high-maintenance structures. You literally built your life around the limitations of local power generation: factories were placed next to rivers, work stopped when the wind died, and most tasks were done by muscle because centralized power was physically impossible. Then came the electric grid. Suddenly, arbitrarily large amounts of mechanical (and later computational) power became available through a tiny wire for pennies. The old model didn’t slowly shrink — it was obliterated almost overnight: Tens of thousands of private steam engines and water wheels were scrapped. Entire industries built around “power proximity” (textile mills on rivers, etc.) collapsed or moved. Human and animal muscle, which had been the primary power source for all of history, became economically irrelevant for most tasks within a single generation. No one missed the “jobs” of stoking boilers or breeding draft horses. The old power structures weren’t reformed; they were made obsolete by the removal of latency and cost. Power went from being a scarce, local, high-friction resource to an abundant, global, near-zero-marginal-cost resource. AI is doing the same thing to intelligence that electricity did to mechanical power. Every corporation, university, government bureaucracy, law firm, hospital, and school system is currently a giant, expensive “intelligence power plant.” They exist because high-quality decision-making, pattern recognition, synthesis, and prediction used to be scarce and high-latency. We built million-person organizations full of rigid hierarchies, manuals, compliance departments, middle managers, and 4-year degree requirements as the only way to concentrate enough intelligence in one place to get anything complex done. Now intelligence is becoming an on-tap utility with ~zero latency and approaching-zero marginal cost. The implications are brutal and fast: A 100,000-person consulting firm is the modern equivalent of a factory with its own coal-fired steam engine in the basement. Most of the S&P 500’s market cap is “legacy intelligence infrastructure” that is about to be worth approximately what private power plants were worth in 1950. Entire categories of white-collar work (and the org charts built around them) are draft horses watching the first locomotives roll by. The phase change isn’t “AI gets better.” The phase change is when intelligence stops being a capital-intensive, geographically constrained, human-speed resource and becomes something you pull from a wall socket. Every structure that evolved to route, concentrate, credential, or throttle intelligence because it was scarce becomes pure drag. We’re not ready for how fast the old “intelligence power plants” get decommissioned once the grid is fully live. Most of them won’t downsize — they’ll just cease to have a reason to exist, the same way no one decided to “reform” the industry of maintaining 19th-century boiler rooms. The plug simply got cheaper and better. That’s the real story: the greatest demolition of institutional infrastructure in history, triggered by the removal of scarcity in the one resource every modern structure was built to husband.

The pez dispenser design is nuts. So incredible to watch.

The correct answer to “are the models conscious?” is “I don’t care because I am on Team Human, not Team Consciousness”

x money gonna shake things up. A lot of forces against big banks now. Days are numbered. Just a matter of time.

With muse the audience and interface just changed. This has massive implications for apps payments etc.

A few thoughts on the current state of venture capital. When the Music Is Playing In July 2007, a few weeks before the credit markets seized up, Chuck Prince, then the CEO of Citigroup, gave an interview to the Financial Times. The line everyone remembers is this one: "As long as the music is playing, you've got to get up and dance." He was mocked for it for years afterward, and he lost his job a few months later. But I have come to think he was saying something honest. He wasn't claiming the music would play forever. He was admitting that he couldn't sit down while it was still going, and neither could anyone else in his seat. I've been thinking about that quote a lot lately, because right now is the most disorienting period in venture capital I can remember, and I have been doing this for a while. Here is what makes it disorienting. It's not that things are bad. Some things are spectacular. We have companies in our portfolio growing faster than anything I have seen in my career, and I don't say that lightly. At the same time, we have companies with no revenue, no product, and a founding team you could fit in a conference room raising billions of dollars at valuations of $10 to $50 billion. Both of these things are true at once, and if you try to reason about them with the same framework you will drive yourself crazy. Two ideas have helped me make sense of it. Neither is mine. The first is reflexivity, which George Soros has been writing about since the 1980s. In most of life, perception follows reality: the weather is what it is, and your opinion of it changes nothing. In markets, it runs the other way too. Prices change what participants believe, and what participants believe changes the prices. The feedback loop can run for a long time, and while it's running it looks exactly like progress. Here is how reflexivity is playing out in AI. Full disclosure: Menlo is an investor in Anthropic, so read the following with that in mind. People watched a frontier lab go from a $4 billion valuation to $18 billion, then $60 billion, then $180 billion, then $380 billion, and now something close to a trillion. They drew the obvious conclusion: that is what a neo lab looks like. So the next neo lab gets priced off that path, not off anything it has built. Then it gets marked up in a subsequent round, and the markup itself becomes the proof. Look at Thinking Machines. Look at Reflection. At that point valuation has stopped being an output of the metrics and has become the metric. Nobody is discounting cash flows. They are discounting the last round. Soros is very clear about one thing, and it's the part people skip: you cannot know when or how a reflexive process ends. You only know that it does. Every one of them has. The second idea is Chuck Prince's, and it explains why smart people keep dancing even when they can see the loop for what it is. As far as I can tell, there are two groups on the dance floor. The first group got in early. Firms like ours were in some of these AI companies before the numbers got silly, and the paper gains are enormous. When you are sitting on gains like that, you start to feel like you're playing with house money. I have been around long enough to know that house money is the most dangerous kind, because you don't respect it the way you respect money you had to earn. The second group missed the early rounds and knows it. Their LPs know it too. So they are trying to make up for lost time by writing very large checks very late, which is the one strategy almost guaranteed to turn a missed opportunity into a real loss. House money on one side, FOMO on the other, and reflexivity feeding both. That's the whole story. Everyone has a reason to keep dancing, and the reasons are different, which is why nobody can talk anyone else off the floor. So what do you do? The instinct in our business is to answer with company identification: just pick the right neo lab and you'll be fine. I think that's the trap. When price has become the signal, being right about the company is not enough, because you can be right about the company and still be wrong about the price by a factor of ten. The public-market investors I admire figured this out a long time ago. They spend as much time on how much to own as on what to own. The winners in venture over the next decade will be the firms that treat portfolio composition and position sizing as seriously as they treat sourcing. How much of the fund is in companies whose valuation rests on the last round rather than on revenue? What happens to the portfolio if the reflexive loop breaks next year instead of in five? Those are not exciting questions. They are the ones that will matter. The music will stop. It always does. Dance if you must, but know where the chairs are.

"In markets, it runs the other way too. Prices change what participants believe, and what participants believe changes the prices. The feedback loop can run for a long time, and while it's running it looks exactly like progress."

Software should be fun. It should make you feel something, positive hopefully. Utilitarian software is the worst. The human experience is why life is worth living (even for something as teeny tiny as a terminal).

Agents.md is just documentation with hope attached.

Looks like today may be a record day for token volume % of open models on Vercel AI Gateway: 🟦 Open 78.4% 🟨 Closed 21.6% While spend 💲 usually tells a different story, #3 and #4 today are Moonshot AI & DeepSeek. Adding Z⁠.ai, their combined spend surpasses OpenAI (#2). (Do note that's the spend for inference of the model across providers (mostly in the US), not revenue going directly to the open weight labs.)

Another awesome report from GOAT @mjmauboussin and @dcalz... “Free cash flow (FCF), profit minus investment, determines long-term shareholder value. Assuming a company can finance its operations, negative FCF is not only fine, but desirable, if returns on investment exceed the cost of capital.” “The point of emphasis is that negative free cash flow is perfectly fine provided that ROIC remains above the cost of capital.” “One result of her [Victoria Dickinson] method is that companies can switch stages based on their patterns of profitability and investment spending. This includes going from a later stage, such as maturity, to an earlier one, such as growth.”

The "whiteboard defense:" I should be able to pull you aside at any moment and ask you to explain any customer-facing system you've shipped. You should be able to clearly explain how it works and defend the decisions you made. This is my benchmark for responsible AI usage. I don't expect line-level familiarity with the code. I don't care if you remember the exact function name or implementation detail. You may not even know it. I don't care. But if I ask "why did you do X instead of Y?", "what happens if this actor behaves maliciously?", "what data structure did you use here and why?", or "where does this fail?" you should be able to answer confidently. For PoCs, demos, experiments, whatever: I don't care. Generate 100% of it and understand none of it. Speed over quality every time in those specific scenarios. But if you're shipping customer-facing work, you can't be shipping things you don't understand at a high level.

Easy to predict. What is not showing up yet is the shift to OS models which is happening. Fasten your seatbelt. The question remains around the price elasticity of demand.

I’ve decided it’s time to have a little more fun on X this year! 😄⚡ I get to speak at some amazing events, meet fascinating people, hear great stories, and occasionally find myself in places I never expected to be. So I figured… why not share some of those moments here? And there’s more! I’m also excited to launch my new merch. A little Woz spirit, a little fun, and hopefully a few things bring a smile to your face. This is just the beginning. More adventures, more stories, and more surprises to come! WozMerch.com

One thing a lot of (tech) investors don't seem to understand about raising rates is that asset prices declining is one of the primary ways the tactic proves effectic- the Fed may well need to keep raising until asset prices decline. It works like this: 1) The Fed needs to cool inflation by cooling the economy. 2) Its main lever is that it can set the risk-free rate of return. 3) To cool the economy, it increases that risk-free rate of return, which means investors should require higher returns from other investments (i.e. capex), which will make some investments uneconomic 4) All else equal, higher required returns = lower asset prices and less investment in something, which cools the economy. That isn't the only effect- higher rates can also cool consumer spending (because goods are financed, so higher rates = less affordability), but generally speaking, investment is more volatile than consumer spending and tends to go first. There are obviously various types of investment going on in the economy (building real estate, building data centers, building factories, building power plants etc.), and typically one (or all) of them need to cool off a little for an increase in rates to have the desired effect on inflation. So there really isn't a scenario where raising rates does the job and doesn't negatively impact some part of the economy.

Just do the reps. "How did they go from the least-defensible to the most-defensible major tech company in ~30 years? Relentless execution and reinvestment. Their moat is 7,500+ days of building things, taking the profits, and building more things."

X is the best party on the internet. Yes, there's some shady shit going on over there in the alley, but in the main room? People having fun, sharing technology, and connecting like never before. X is an essential part of Omarchy's success. $42B was a bargain!!

Well thought out.

There's lots of talk about the tokenpath, but it has never been more important for startups to also be on the tweetpath. I've been discussing this with various folks and I know it sounds crazy/the result of being a terminally online VC, so here's why it matters (and how to do it): First, many of today's great companies (@tryramp, @cognition, @modal, @openAI, @anthropic, @fal etc.) have cultures where a significant share of employees (and the founders) are regularly updating the world on what's going on here. It's hard to argue this is working against any of them, and it's also very different from history. I can't remember a single SaaS company that had this culture. So something has changed- and without a doubt, it is the pace of innovation in the fundamental primitive products that underpin what it means to be a startup today. SaaS was about blocking and tackling on top of cloud infrastructure building blocks (e.g., S3, EC2, etc.) that are still recognizable almost two decades later. In AI, new models drop on a weekly basis that have the potential to transform the value proposition of a product. Astra is a step-function change in computer use. H3 Max is a step-function change in text-to-video speed. Both dropped in the last few weeks (!!). For better or worse, X is the platform on which news and analysis about these releases propagates. You can literally have access to @thsottiaux's commentary around Astra adoption, @Yuchenj_UW's read of model tradeoffs, @burkaygur's generative media insights, etc. in real time, and if you don't have them, you're behind by months in an era where months really matter. This can determine product decisions, spawn partnerships, etc. That's an argument for scrolling, but there's an equally compelling argument for participating. For a start, this critically important conversation is happening here and it is better to be a part of it than apart from it. It also feels/is fundamentally different than the navel-gazing that dominated X of yore- because almost every day there's a new thing to talk about/analyze/contribute. I used to have flashes of annoyance when I saw VCs dip into politics- don't see much of that on the timeline anymore, because there's so much to talk about in tech. Because of that, there's an opportunity to really express/brand oneself and build up a reputation/point of view that arguably didn't exist before and is way more valuable now. I know startups that have gotten acquisition offers, employees who have been poached to amazing companies, etc. all because they demonstrated progress or competence here (in part by showing that they are in the top 1% at consuming/processing the aforementioned constant information flow). The gap between employees who are on the exponential and not has never been larger, and this is the best/most observable way to demonstrate that you are on the exponential... and yes, exactly the same logic applies for companies too, and then it gets recursive, because employees who are on the exponential want to work for companies who are on the exponential, and so on. And that's the last argument- the effort of demonstrating that you're on the information exponential (by participating in conversations that feature social risk if you haven't kept up) will necessarily help you stay there while inertia/human nature fight to drag you down. That alone is mega valuable on top of everything else. You will learn more, faster, if you force yourself to be engaging with folks on the frontier. It isn't the distraction it used to be- it is a learning accelerant. And keeping your head down/locking in/plowing forward on a strategy without staying abreast of the news of the day is less advantaged approach today than perhaps at any time in human history. If I've convinced you- here's my advice on how to kickstart this as an individual/company. As an individual- post about cool stuff you're doing or learning about, get into a habit of thinking about what other people might find interesting/incremental and then send it and never think about it again unless someone replies. I'm writing this in one fell swoop and won't even reread it. If there's something you want to understand better, odds are someone else is thinking about it too. The bar is lower than you think- and if it catalyzes you to learn/sharpen up on an important topic, all the better- but you first have to get past the fear that makes you edit a post 20x before sending. That's a time suck for little incremental benefit. As a company- for one, the culture needs to encourage and promote that employees sharing neat things they've done is good for the aura of the company. Ramp Labs under @alexstauffer_ posted a blog post on Opus playing Roller Coaster Tycoon that I still think about as a canonical example of org culture marketing. Companies need to grok that this isn't a distraction; it's core- and founders can really lead by example here. I've heard case studies where employees post their wins to X because they have a hunch that the CEO has a better chance of seeing it here than on Slack... and I don't think that's a cultural bug per se. The only real rules are: there can't be a super onerous process for getting posts approved (high trust), it works best as a mafia of people supporting each other, and, of course, one has to be doing/learning cool stuff that relates to their business problems. If you don't have that, you have another set of problems. Very curious if others have thoughts/pushback here! That's part of the point, after all

Give it another 6 months for it to gain widespread adoption.

Let's fix 3 logos. Which fix do you like best? Switch 2, U-Haul or Jet Financial?