chari

Vigilance Decrement: "FSD works almost all of the time and that’s the problem: A machine that constantly fails keeps you sharp. A machine that works perfectly needs no oversight. But a machine that works almost perfectly? That’s where the danger lies." theatlantic.com/magazine/2026/…

20 years after Youtube was bought by Google, its ad revenue exceeds that of Disney, NBC, Paramount, and Warner Bros Discovery (WBD) combined! techcrunch.com/2026/03/10/you…

Pretty amazing

@doristsao Wow amazing! I assume you’ve had many years’ experience..

including yours truly

@JJEnglert Guide

I've used Claude Code to build 20+ projects in the last 6 months. Thousands of new users across them. And I've never written a single line of code. I just dropped a 24-min video with my top 10 tips for non-developers — the exact playbook I use every day to run multiple AI agents that handle work that used to take me a full week. This is the best beginner guide to learning and building with Claude Code out right now. Every tutorial I found assumes you're a developer. This one doesn't. I cover everything from first install to running multi-agent workflows — with live demos and real examples for every single tip. How I set up new projects, how I got Claude to match my writing style, how I automate repeatable workflows with one command, and how I run multiple agents working on different tasks at the same time. I also built a full resource repo to go alongside the video — curated video tutorials, the best skill libraries, plugin directories, MCP server guides, written docs, community links, and a starter CLAUDE.md template you can copy-paste into your first project today. Comment "GUIDE" and I'll send you the full guide with everything you need to learn Claude Code! (make sure we're connected so I can DM you)

Prof. Richard Feynman talks about algebra.

Ominous

@realroseceline Hence the argument for drone-based delivery. No need to navigate through traffic either, just get from A to B in a straight line, with minimal fuel consumption and drop the package off on the doorstep or back yard

I think a small, stripped down vehicle built only for deliveries would be a huge positive for food delivery companies. Not because robots are cool, but because most of what makes cars expensive is totally unnecessary for food delivery. You don’t need seats, mirrors, speakers, etc. You basically need a locked box on wheels with sensors and software. It would work more like a robot than a car. It drives to the restaurant, pings them, they put the food into a compartment, and it drives to the customer. The customer gets a code in the app, opens the compartment, takes the food, and that’s it. The reason this matters is simple. Paying the driver is one of the biggest costs in food delivery. Labor doesn’t scale well, but autonomy does. I haven’t fully modeled every number, but it’s hard to imagine this not helping margins in a big way if utilization is high. 🌹

great post by @stevesi, applying a broader historical lens to technology transitions, such as the current one, to suggest that there will be more need for software than ever before

It turns out that all those awkward and embarrassing moments we suffer through during the course of life are simply due to the absence of background music.

The Ukraine war is leading to a number of innovations in drone tech: autonomous targeting (to evade wireless jamming attacks on the controller comms link), visual positioning for GPS jamming resistance, drone swarms to maximize shots on goal.. nytimes.com/2025/12/31/mag…

Thought-provoking essay on what knowledge work might look like when agentic workflows run continuously, "delegating busy work to minds that never sleep" notion.com/blog/steam-ste…

@levie Doesn't the Jevons paradox rely on highly elastic demand, where the volume of work expands enough to offset productivity gains and avoid net workforce reductions? This may not apply to all jobs - e.g., tax preparation, where demand may be more inelastic..

Decoding India's air emergency The fundamental issue is not just the pollution itself, which is a kind of constant, but also the stagnation and the lack of ventilation. The winds provide horizontal ventilation. Vertical ventilation is what is actually creating the entire hazardous condition when people are suffering and going to the hospital: @sri_srikrishna #5Live @Sonal_MK

This is the EV flippening. Electric vehicles now fill up roughly as fast as gas. The new BYD chargers add 400km of range in just five minutes.

This is wild! Johannes Schmitt used GPT5 to solve his own open problem on intersection numbers on moduli spaces of curves (the proof turns out to be unexpectedly simple, "low hanging fruit"). He wrote up the paper, being careful to point out which *entire paragraphs* were written by a human vs AI. Will this become the future protocol?..

@gothburz Hilarious and genius. Thank you for the insights

Last quarter I rolled out Microsoft Copilot to 4,000 employees. $30 per seat per month. $1.4 million annually. I called it "digital transformation." The board loved that phrase. They approved it in eleven minutes. No one asked what it would actually do. Including me. I told everyone it would "10x productivity." That's not a real number. But it sounds like one. HR asked how we'd measure the 10x. I said we'd "leverage analytics dashboards." They stopped asking. Three months later I checked the usage reports. 47 people had opened it. 12 had used it more than once. One of them was me. I used it to summarize an email I could have read in 30 seconds. It took 45 seconds. Plus the time it took to fix the hallucinations. But I called it a "pilot success." Success means the pilot didn't visibly fail. The CFO asked about ROI. I showed him a graph. The graph went up and to the right. It measured "AI enablement." I made that metric up. He nodded approvingly. We're "AI-enabled" now. I don't know what that means. But it's in our investor deck. A senior developer asked why we didn't use Claude or ChatGPT. I said we needed "enterprise-grade security." He asked what that meant. I said "compliance." He asked which compliance. I said "all of them." He looked skeptical. I scheduled him for a "career development conversation." He stopped asking questions. Microsoft sent a case study team. They wanted to feature us as a success story. I told them we "saved 40,000 hours." I calculated that number by multiplying employees by a number I made up. They didn't verify it. They never do. Now we're on Microsoft's website. "Global enterprise achieves 40,000 hours of productivity gains with Copilot." The CEO shared it on LinkedIn. He got 3,000 likes. He's never used Copilot. None of the executives have. We have an exemption. "Strategic focus requires minimal digital distraction." I wrote that policy. The licenses renew next month. I'm requesting an expansion. 5,000 more seats. We haven't used the first 4,000. But this time we'll "drive adoption." Adoption means mandatory training. Training means a 45-minute webinar no one watches. But completion will be tracked. Completion is a metric. Metrics go in dashboards. Dashboards go in board presentations. Board presentations get me promoted. I'll be SVP by Q3. I still don't know what Copilot does. But I know what it's for. It's for showing we're "investing in AI." Investment means spending. Spending means commitment. Commitment means we're serious about the future. The future is whatever I say it is. As long as the graph goes up and to the right.

@doristsao @NikoMcCarty @AsteraInstitute Totally! Would be great to catch up. Maybe we can get Sri to join also

@anchari @NikoMcCarty @AsteraInstitute Thank you, Chari! We should get together one of these days

I am deeply excited to share that I will be joining @AsteraInstitute to start a new effort to understand how the brain generates consciousness and intelligence.

@hsu_steve @awaken_tom lol “strategists”

"neoconservative plan = the idea to blockade China just long enough to reach superintelligence, then use SI to cripple China and secure world domination." This is correct, as stated explicitly by US strategists and also Dario Amodei, Eric Schmidt, etc. However at least Eric has figured out that this isn't going to work as initially envisaged. There was a wildly wrong estimate of the "just long enough" timescale, as well as of PRC capabilities such as 1. building SOTA models relying on algo optimization rather than brute force compute 2. building AI chips and semiconductor supply chain. Most neocons overestimate US capability and underestimate PRC, Russian, etc. capabilities. This led to the Ukraine disaster, and also now the risk of losing the AI race.

Watch this debate - relevant to H200, PRC vs US AI and semiconductor race. Discusses "sprint to AGI" justification by Biden-era team for chip and semiconductor controls that ultimately solved the "alignment problem" for the PRC semiconductor ecosystem and (in my view) hurt the US in the medium to long run. I think most AI researchers now accept that AGI = self-improving AI, or something that could lead to an accelerating science+tech advantage for the US over PRC, is further in the future than was thought by naive scaling proponents back in the Biden era. It's a longer race now, probably extending to the 2030s, with infra = power grid, datacenters, etc. playing as big a role as single-chip capability.