dv’s 2nd Account

My conversation with Tony Xu (@t_xu), co-founder & CEO of @DoorDash. 0:00 DoorDash MVP in 43 Minutes 1:39 How Delivery Worked in 2013 3:17 Small Business Roots and Insight 5:48 Why Restaurants First 8:24 Palo Alto vs San Francisco 11:03 Early Customers and Unit Economics 15:22 YC Summer Three Questions 19:50 The Hidden Complexity of Delivery 22:02 Competing on Invisible Details 23:54 Chaos Data and Experiment Loops 30:58 Trust Reset Every Day 31:30 Stanford Game Meltdown and Refunds 34:41 Scaling Through Experiments 37:37 Customer North Star Metrics 40:10 CEO Customer Support Habit 42:55 Anecdotes Versus Data 46:52 Eternal Mission Local Economies 50:09 Turning Data Into Merchant Growth 59:12 New Products Beyond Delivery 1:01:14 Autonomous Delivery Strategy 1:05:06 Hiring Rhodes Scholar Navy SEALs 1:12:46 Driver Switch Experiment 1:13:42 Who Delivers and Why 1:15:33 Hiring for Action 1:18:07 Earned Secrets via Experiments 1:20:01 Money vs Problem Solving 1:21:18 Thousand Days of Hell 1:26:04 Staying Sane as CEO 1:30:07 Ignore the Stock Price 1:31:44 Two Operating Systems 1:35:17 Internal Venture Stage Gates 1:38:17 Learning from Founder Peers 1:42:29 Jiu Jitsu Lessons 1:44:37 AI Changes the Loop 1:47:01 Data Needs Action 1:48:24 Closing Thoughts Includes paid partnerships.

The XR Week Peek (2026.03.30): Meta to launch two new smartglasses, the metaverse dies again and again, and much more! skarredghost.com/2026/03/30/met… #VirtualReality #Meta #smartglasses

Ollama is now updated to run the fastest on Apple silicon, powered by MLX, Apple's machine learning framework. This change unlocks much faster performance to accelerate demanding work on macOS: - Personal assistants like OpenClaw - Coding agents like Claude Code, OpenCode, or Codex

prediction: AGI will have a harness, but it will be just-in-time, self-assembled based on the current Task+context today our human priors dominate the details of harness engineering. we inject our notions of good engineering practices into the harness design but systems like alpha-zero have shown that this can be a local-minimum, models will get better than us at this eventually it’s good to lean into more model autonomy while today leaning into good harness design using our priors we want to get work done today! that means lean into the frontier of using agents to do good work today which is practically a lot of context + harness eng at the same time we can tinker at the frontier of the future by giving models way more autonomy

I’ve used Opus 4.6 and GPT 5.4 on a mix of projects since release, and want to break down where I think they uniquely excel. It’s more nuanced than you’d think! Rigor of code - GPT 5.4. It goes the distance validating its work without asking. Opus needs explicit instruction to do this, and even then, it misses more edge cases. Clarity of code - Opus 4.6. Claude is a better communicator, which carries into the code. Variable names are clearer and less mechanical, which improves reviewability. This is very important since code review is the bottleneck for most engineering teams. It also adds the right amount of doc comments. GPT simply never comments or explains its work; it’s like working with an obtuse engineer that wants the solution to speak for itself. Sometimes it does, other times not. Similarly, rigor of plans goes to GPT 5.4, while clarity of plans goes to Opus 4.6. An interesting point though: GPT performs better talking through a strategy without a plan, while Opus needs planning mode to put in any rigor. I find myself forgetting plan mode altogether using GPT 5.4. Quality of research - toss-up. Opus spends longer researching with web search, but GPT spends longer studying the existing codebase. You may think codebase research matters more, but researching how others solve the same problem can be just as important. Maybe more important for greenfield. Quality of conversation - Opus 4.6. It’s just better to talk to, which matters using these things everyday. GPT 5.4 was clearly trained to challenge the user more, which results in a tendency to *always* say you are wrong. I’ve had bizarre interactions where GPT claims something is “not quite right,” the restates exactly what we’ve decided on in the last turn. On a personal level, it’s annoying. On a practical level, it makes iteration on a plan slower. THAT SAID, it takes sufficient pushing for Opus to challenge your thinking in this way. Simply say “I’m impartial” and ask questions to avoid that, as you would a person. Overall winner - Opus to make it work, GPT to make it good. I don’t have a good system of when to switch tools, but on average, I prefer Opus early on and GPT for optimization and discussing architectural decisions. Opus is also better for any design related tasks (but state management in frontend apps is better handled by GPT).

Genspark Claw can now make calls for you. Just say “make a call for me” and try it. Book a table, check store hours, or ask about appointment availability. Try it and let us know what you use it for 👀 genspark.ai/claw

Tim Ferriss (best known for his “4-Hour” book series) on how he uses AI: “I hesitate to use AI for anything I want to keep in my head.” Because AI doesn’t just assist — it can fully replace your thinking. The cost? Cognitive muscles atrophy fast.

I'm literally just building every idea I had today instead of just putting it into Apple Notes Open @conductor_build, create a new project, open a new branch, start on it... and it will be in reality shortly.

So many PR's to land tonight for GStack. The community is amazing and giving me so many good ideas and fixing bugs. Thank you to the #gstackfam

bep is a crazy handle

RT @bryan_johnson: When I started Don't Die 2021, we evaluated all the scientific evidence for the most powerful anti-aging therapies. Psyc…

Embrace boiling the ocean

I swear we didn’t slow it down. It really is just that slow… 😅

We've gotten questions about how Phota works, so we wanted to share more about what we've built. Phota, like many modern AI products, is a system of multiple models, each optimized for a different part of the generation pipeline. Our proprietary model focuses on identity, which enables us to generate and edit photos of real people and pets. For base image generation, Phota uses leading foundation models (both open and closed source - including Nano Banana). On top of those, we've trained our own identity model on both in-house data and user-uploaded photos to preserve identity consistently. That identity layer is the core of Phota's differentiation. Phota has always been built as a multi-model system; we have not presented it as a single foundation model trained end-to-end from scratch. From the beginning, our core research effort has been focused on identity and photography. Phota was built by former Adobe researchers who have spent years working on image models. Our launch this week focused on the product experience and the outputs - but we've also heard the interest in going deeper on the underlying research, and we're excited to share more over time. We invite you to try Phota yourself, and we're eager to hear your feedback.

Project Hail Mary

tech bros after building a personal CRM

Mixed Reality is no longer a gimmick

Claude's private thinking steps. Reacted exactly like a shocked human reading the morning news. Someone asked Claude a question about Iran. Claude’s extended thinking discovered the Iran strikes mid-response. The vibes shifted immediately It reads the first search result and thinks, "Whoa." that’s not a human reacting to the news, that is the actual, unedited internal thought process AI caught off guard. Then, it searches specifically for the airstrikes to confirm, and its internal monologue literally says, "Holy shit." --- reddit .com/r/ClaudeAI/comments/1ribnke/claudes_extended_thinking_found_out_about_iran_in/

Today's @symbolica harness is a clear example of what human-crafted targeting can achieve on ARC-AGI-3 public demo set You can "buy" performance with benchmark-specific prompts/strategies Their approach could still contain useful ideas, excited to see what the community finds

We have first ignition on the Gen 4 propulsor (and blessings by the good luck falcon)