Andrew Diaz

Engineer | Founder: Building CV infra for Pro Sports

[Andrew Diaz @ADiaz_NoLimit]

Renaissance 2026 -❓

[Andy Konwinski @andykonwinski]

Introducing Headlong, an open source microharness for persistent agents: self-guided agents that think continuously.

Most agent harnesses are reactive: you send a task, the agent completes it, and then it sits frozen until the next request. Cron jobs and heartbeats wake it up to run a checklist and put it back to sleep.

A Headlong agent is never asleep. It keeps generating thoughts about whatever it decides is interesting, in a self-guided loop inspired by human inner monologue. Your message doesn't start a session. It's one more observation that lands in the agent's thought stream, and the agent decides if and when to reply.

Headlong is built on the idea of persistent agency: continuous inner thought generation between external interactions. The agent sets its own interests and priorities, comes up with its own projects, and sometimes pings you unprompted with progress.

To keep our prototype as simple and small as possible, we implemented Headlong as a microharness: a complete agent harness in under 10K lines of Bash, organized as a handful of small executables. It includes a loop that generates the next thought, shellm (a recursive language model written in Bash), a trajectory stored as a DAG of jsonl files, and context as a projection of that trajectory.

We've been running one Headlong agent internally at Laude for several weeks. The whole team talks to it over Slack and Telegram, and every conversation lands in its single stream of thought. It works in its own fork of Headlong and we've pulled over 50 of its commits into main.

One night, with nobody talking to it, it went back to check whether a recall process it had built was actually wired into its mind, found that it wasn't, diagnosed and fixed the bug, and verified the fix end to end. 48 minutes, no human asked for the fix or was in the loop at any point. Every step is a timestamped line in its log.

Things broke too, and we wrote those up. Background thinking costs us $1 to $2 an hour, our agent stopped its own service three times by accident, and self-delegation died on day one. Details in the post.

One line installs everything and starts an agent. Use a dedicated sandbox and spend-capped API key; it runs real shell commands and thinks around the clock.

Headlong is research software, be careful!

curl -fsSL https://t.co/P2Q5Ch8C4e | bash

Launch post: https://t.co/HhWULUTkpH

Repo: https://t.co/U4JunpSTYt

Headlong is a @LaudeInstitute / MIT collaboration.

[DAN KOE @thedankoe]

When calories became cheap, convenient, and abundant, bodies decayed.

Thinking is now cheap.

[Yann LeCun @ylecun]

@firesidealpha Sudden Clarity Sam realizing that real-world dynamics equations have inertia and friction terms.
(Particularly when they concern large systems like human societies).

If only the singularitarians, AI doomers, AGI-is-near crowd, and other SV cultists could realize that....

[Andrew Diaz @ADiaz_NoLimit]

I'm not dropping my claude subscription anytime soon. No sir, Anthropic products are the nuts!

[Gavin Baker @GavinSBaker]

Good chart from @TMTBreakout

July accelerated both MoM and YoY for the sum of OpenAI and Anthropic.

Some chance Anthropic shifted from gross to net reporting for ARR. Meaning the 65b is a more conservative metric than the $47. Smart thing to do before an IPO.

Openrouter data shows that AI broadly has accelerated further over the last 3 weeks. Open source taking share is positive for AI infrastructure demand.

[Andrew Diaz @ADiaz_NoLimit]

Choose your VLMs wisely

[Jitendra MALIK @JitendraMalikCV]

Scientific terms should have precision. If we use the terms VLM, VLA, WAM in an indiscriminate fashion, as is becoming common in robotics, we are not helping clarity in communication. Let's keep the historical origins of these terms in mind. VLMs arose as multimodal extensions of LLMs-the training was for tasks like VQA (VIsual Question Answering). These capture the static semantics of the scene behind an image. No dynamics. World Models (e.g. @ylecun , Ha & Schmidhuber 2018) on the other hand are primarily dynamics models, which go back to control theory -1960 (Bellman, Kalman etc.) This makes them natural for robotics planning / policies- I am in a state s, what action a should I perform to get to state s'. In classical control, these models were written down a priori by modeling the physics of the system; today we think of them as learned neural networks trained from temporal data e.g. video, robot trajectories. But the concept is the same. We shouldn't mix this concept with VLMs.

[Brendan (can/do) @BrendanFoody]

Prediction: a frontier lab will launch/relaunch a fine-tuning product in the next 3 months to compete with the RLaaS companies

This feels inevitable with the severity of compute bottlenecks.

[Joon Sung Park @joon_s_pk]

I’ve always thought of simulation as painting a portrait. A great portrait helps you understand its subject beyond what’s visible on the surface.

This new scaling law will let us paint the highest-resolution portrait of our society yet. That’s worth getting excited about.

[Garry Tan @garrytan]

Reminder: the datacenter discourse is a documented psyop and they’re winning

[Andrew Diaz @ADiaz_NoLimit]

Wait, but she doesn’t have purple hair 😱

[Chamath Palihapitiya @chamath]

Here’s the ad: https://t.co/wPcbazTUdF

[Billy Binion @billybinion]

Fascinating. Some data centers came to a struggling town in Washington state. Did it run out of water? Was the community destroyed?

Let’s see. The town…built a new high school, hospital, library, sewage systems, and police & fire stations. Poverty fell from 29% to 6%.

Oh.

[Daniel San @dani_avila7]

Quick demo of how I use session to session messaging in Claude Code + Ghostty

(Save this and try it in your own workflows)

1- I always start with a regular session and a panel running claude agents

2- The first session becomes the orchestrator. It creates the initial context, then distributes the work across other sessions using forks

3- Each fork starts with that context + a prompt telling it what to work on

4- I rename each session with ctrl + r to keep things organized and make it easy to message them from the main session

5- Using @ I send messages to the sessions that are running and ask them to report their task status back, so the main session always knows what everyone is working on

This way, I only need 2 or 3 panels open and can switch between tasks as needed

The days of opening endless terminal panels and running Claude Code 100 times just to look productive are over!

What we need now is ORCHESTATION
one main session coordinating everything through session messaging + Agent View

And the key is context, build it in the main session first, then fork from there

That’s what makes the entire workflow click

[Justine Moore @venturetwins]

You guys sure you don’t want more data centers?

[Naveen Rao @NaveenGRao]

The single biggest thing that the general populace is missing about AI is that it will create 100x more entrepreneurs. People focus on job "loss" (partially the fault of AI labs' rhetoric)...it's not loss, it's change. AI enables human agency and enables more people to deliver value! It seems so obvious

[Michael Romanenko @romanenko_m]

@reillyjodonnell Yes! Extremely excited for what’s @zeddotdev are cooking with DeltaDB. It does sound exactly how this would look like in Syntax’s latest episode (https://t.co/9SPH0yUpey).

[Vinny Lingham @VinnyLingham]

I’m getting a whole bunch of Chinese subscribers (presumably AI bots), that pay the subscription and ask me to follow them back. Anyone else seeing the same thing?

What are they trying to do?

[Andrew Diaz @ADiaz_NoLimit]

Time between model releases have gone exponential. The implications are impossible to fathom. We are in the singularity.

[Elon Musk @elonmusk]

@beffjezos Grok 4.7 is significantly better than 4.6 and should be ready in 3 to 4 weeks.

Initial training is complete and now we’re adding a massive amount of SpaceX company data in supplemental training. This will be something special.

[Jack @jackunheard]

Let this sink in....

15 women... gathered together... in a room... for hours...

and they couldn't decide on what a woman is.

[a16z @a16z]

Travis Kalanick says entrepreneurs need pissed-off energy:

"When you get into your 3rd and 4th, this may be even my 5th at this point, company, you get fricking really good at doing stuff."

"Things that used to take me a day or 3 days, a lot of stress, anxiety, like how am I gonna do this, take me like 45 minutes."

"I'm so used to adversity at this point that it's kind of normal, so I don't get as pissed off. But being pissed off makes you good as an entrepreneur, so I gotta make sure I'm still fired up."

"You get too used to adversity, and that's kind of a problem."

@travisk @bhorowitz

[SkalskiP @skalskip92]

here are 10 strongest VLMs in my vision benchmark, sorted by cost

Gemini 3.7 Flash is insane

- 2nd in object detection (behind Qwen3.8-Max)
- 2nd in text extraction (behind Gemini 3.5 Flash)
- 2nd in image reasoning (behind Gemini 3.5 Flash)

- over 3x cheaper than Gemini 3.5 Flash
- over 2x cheaper than Qwen3.8-Max

benchmark: https://t.co/RK74sAFJDt

[Andrew Ng @AndrewYNg]

New: A map of the most important skills in AI Engineering. https://t.co/VVkn1Dqp1N