brotchie

GPT 5.6 already is an explainer agent.

Fork Tao’s convo and prompt this (with your own math level described).

GPT did a great job of translating Tao’s questions and concepts (e.g. “pre image”) into a progression I could understand.

“Ok I have a PhD in financial math and undergrad in engineering math. I have almost zero knowledge of polynomial algebra / geometry, I know what a polynomial is and what roots are but not much beyond that. Could you try and explain to my level what questions the user I the conversation has asked and what the agent has responded with, we can probably go user query by user query to build up”

My hand wavy math.

One person dedicated to a single project / task baseline 100% total output.

One person split between two tasks / projects that require context switching, 70% total output at best, perhaps 50% total output, a loss of 50%.

One person working on a single task that they’re passionate about / intrinsically motivated to complete: 150-500% output.

When you have one task / project to think about and focus on, it engages your brain: sleeping, showering, walking, your subconscious is processing.

Adding just one extra project / task that requires a context switch mostly kills this subconscious processing mechanism.

3000 RPM is 50Hz right?

I said a better “human driver.”

Waymo is probably safer, but safety comes at the risk of not behaving like a human, and hence causing confusion for other human drivers.

Tesla drives more like humans drive.

99% of my FSD disengagements are navigation related: leaving it too late to merge into an exit lane (or CA drivers speeding up to not let me merge), not being aggressive enough at going into the opposing lane to overtake a stationary truck (though, it's a lot better at this in recent versions).

I can only recall one "oh shit" moment on an older FSD version when it carrier too much speed down a steep hill in SF and went through a 4-way stop and I had to slam the brakes. A lot of the strangeness / hesitance of previous FSD version is gone in the latest updates.

I don't expect self driving to be 100% perfect. I feel like FSD today is a 90th percentile human driver with super human reaction speed + visibility. Holding a "never gets into an accident" it too high of a bar. FSD consistently drives better than me (with the caveat of navigation: I know the city, I know which unprotected left turns are easy vs. hard, I know which routes tend to be faster due to local conditions, etc).

FSD is excellent in how reactive it is to thing you don't even see: e.g. car on the 101 at slightly drifts into my lane, I didn't notice it, FSD reacted. I feel WAY more unsafe now riding in an Uber vs. hands off FSD'ing in a Tesla.

I've done 234 Waymo rides and ~5k miles driven FSD in a Model Y (HW4), all in and around the bay area.

The gap isn't that big. Tesla still needs supervision (most around navigation honestly). Waymo's have definitely done some really dumb things.

Waymo's certainly feel safer but if I had to choose which was the better "human driver" I'd put it on Tesla.

If consciousness didn’t exist, you wouldn’t exist! you’d have no interior experience.

You don't need quality or flexibility for an AI demo.

Hand crafting a voice agent to schedule simple appointments for a barber in San Francisco where the caller is in a quiet environment, is a one day exercise in prompt engineering.

Building a voice agent to schedule real appointments (on a real calendar of a working business) for real customers, for any business type, in any city: significantly more difficult. Real customers can be on a bad cell connection, have background noise, or worse, there's somebody in the background having another conversation.

Building a working agent isn't the hard part of building a real world agent. It's establishing human and offline evals, identifying loss patterns, hill climbing, capturing and processing user feedback, doing hacks to deal with model limitations, learning how an agent-driven conversation has to be subtly different from a real human conversation, and so on.

Try a different translation.

First time I started to read it, it was a slog and I didn’t get far.

Did a bit of research on translations and chose another one (can’t recall the exact translator).

The 2nd attempt’s translation used more contemporary language, which made it much more understandable and got through it.

We need the equivalent of the MEAN / LAMP stackronym for agents.

100% agreed, the "this is what an agent looks like to write" is the wrong pitch for a new agent framework.

The better pitch would be, "this is how easy observability, guardrails, monitoring, deployment, evals, versioning, A/B testing are with our framework." What the agent code looks like is somewhat incidental.

I'm still on the fence about agent frameworks, they have their place, and it depends on the nature of the agent: e.g. "Low latency, return a good enough response in 3 seconds, vs. working for 3 hours on a problem."

BUT, if you boil it down, an agent really is context building, making an LLM call, executing requested tool calls, parsing the final model output, returning it to some frontend. There's extensions like memory, async tool calls, etc, but not THAT complicated from a traditional software engineering perspective.

Everyone seems to want to build their agent framework. But if you're tasked with building an agent, I've found it much easier and more maintainable to just build 1:1 code for THAT agent: most of the abstractions you get from an agent framework purely get in the way and obfuscate core agent logic.

You end up being forced to use the abstractions chosen by the agent framework, which sometimes are a mismatch for what you're actually trying to do.

Closest I've been to losing vision in one eye was creating these 3x chain links for Burning Man.

Naive thought: I could use a large bolt cutter to cut chain links. Started trying to cut a link, felt it was sketchy, went and put on some safety glasses.

Restart cutting (had these bolt cutters with like 1m long arms), apply full force, jaws slip a bit on the chain, jaws bite hard. Chunks of steel fly into my chin and face, metal chunks embedded in chin, cracked safety glasses. Dodged a bullet.

Ended using a small welded up jig so I could stretch the chain and then use angle grinder to cut the chain links. Still sketchy, but no flying metal chunks.

Wish I had a plasma cutter.

This is a good counter argument to the paper, honestly.

Are you saying that, in some abstract sense, that actually pouring the cup may be isomorphic to running a perfect simulation of pouring the cup?

Genuinely curious about your statement that its an illusion / arbitrary distinction, to figure out if there's a gap in my thinking / reasoning. To me there's a clear distinction between the actual thing happening via physical dynamics vs. us (humans) having creating a discretized abstraction (binary computation) on top of that and running a process on that abstraction.

Maybe there's some true computational universality where the universes dynamics are discrete (definitely plausible) and there's no distinction between how a processes dynamics unfold: i.e. consciousness binds to states and state transitions regardless of how they are instantiated. I did use to hold this view , but now I'm not so sure.

It's the difference between:

  a) Actually pouring a cup of water into a pond (layer zero), and
  b) Running a fluid dynamics simulation of pouring a cup of water into a pond (some layer above layer zero).

Originally rejected the paper premise, but I get it now, certainly made me question my belief that consciousness binds to any arbitrary information processing that's of sufficient complexity.

IIUC the author is saying that the human brain is running directly on "layer zero": chemical gradients / voltage changes, while AI computes on an abstraction one layer higher (binary bit flips over discretized dyanmics).

In essence, our brains are running directly on the "continuous" physical dynamics of the universe, while AI is running on a discretization of this (we're essentially discretizing the physical dynamics and to create state changes of 0 -> 1, 1 -> 0).

My currently belief is that consciousness is some kind of field or property of the universe (i.e. a universal consciousness field) that "binds" to whatever information processing happens in our wet ware. If you've done intense meditation / psychedelics, there's this moment when it becomes obvious that you are only "you" due to some kind of universal consciousness's binding to your memory and sensory inputs.

The "consciousness arises from information processing," i.e. the consciousness field binds to certain information processing patterns, can still hold, and yet not apply to AI (at least in its current form): The binding properties may only apply to continuous processes running directly on the universe's dynamics, and NOT to simulations running on discretized dynamics.

Yeah, I think about this a lot.

Those days of grinding on some grad school maths homework until insight.

Figuring out how to configure and recompile the Linux kernel to get a sound card driver working, hitting roadblocks, eventually succeeding.

Without AI on a gnarly problem: grind grind grind, try different thing, some things work, some things don't, step back, try another approach, hit a wall, try again.

This effort is a feature, not a bug, it's how you experientially acquire skills and understanding. e.g. Linux kernel: learnt about Makefiles, learnt about GCC flags, improved shell skills, etc.

With AI on a gnarly problem: It does this all for you! So no experiential learning.

I would NOT have had the mental strength in college / grad school to resist. Which would have robbed me of all the skill acquisition that now lets me use AI more effectively. The scaffolding of hard skill acquisition means you have more context to be able to ask AI the right questions, and what you learn from the AI can be bound more easily to your existing knowledge.

Gambling ads are to Australia what Pharmaceutical ads are to the USA.

OpenClaw is just like any other tool, you need to learn it before its power is available to you.

Just like anything in engineering really: you have to play around source control to understand source control, you have to play around with database indexes to learn how to optimize a database.

Once you've learned it and incorporated it into your tool set, you then have that to wield in solving problems "oh, damn, a database index is perfect for this."

To this end, folks doing flights and scheduling meetings using OpenClaw are really in that exploration / learning phase. They tackle the first (possibly uninventive thing) that comes to mind to just dive in and learn.

The real wins come down the line when you're tackling some business / personal life problem and go: "wait a second, an OpenClaw agent would be perfect for this!"

For the tax thing. I had Claude write a CLI and a prompt for Gemini Flash 2.5 to do the structured extraction: i.e. .pdf -> JSON. The JSON schema was pretty flexible, and open to interpretation by Gemini, so it didn't produce 100% consistent JSON structures.

To then "aggregate" all of the json outputs, I had Claude look at the json outputs, and then iterate on a Python tool to programmatically do it. I saw it iterating a few times on this: write the most naive Python tool, run it, throws exception, rinse and repeat, until it was able to parse all the json files sensibly.

Not enough time, too many projects. Useful projects I did over the weekend with Opus 4.6 and GPT 5.4 (just casually chatting with it).

2025 Taxes

Dumped all pdfs of all my tax forms into a single folder, asked Claude the rename them nicely. Ask it to use Gemini 2.5 Flash to extract out all tax-relevant details from all statements / tax forms. Had it put together a webui showing all income, deductions, etc, for the year. Had it estimate my 2025 tax refund / underpay.

Result was amazing. I now actually fully understand the tax position. It broke down all the progressive tax brackets, added notes for all the extra federal and state taxes (i.e. Medicare, CA Mental Health tax, etc).

Finally had Claude prepare all of my docs for upload to my accountant: FinCEN reporting, summary of all docs, etc.

Desk Fabrication

Planning on having a furniture maker fabricate a custom walnut solid desk for a custom office standing desk. Want to create a STEP of the exact cuts / bevels / countersinks / etc to help with fabrication.

Worked with Codex to plan out and then build an interactive in-browser 3D CAD experience. I can ask Codex to add some component (i.e. a grommet) and it will generate a parameterized B-rep geometry for that feature and then allow me to control the parameters live in the web UI.

Codex found Open CASCADE Technology (OCCT) B-rep modeling library, which has a web assembly compiled version, and integrated it.

Now have a WebGL view of the desk, can add various components, change their parameters, and see the impact live in 3D.

The way I solved this was that my open claw doesn't interact directly with any of my personal data (calendar, gmail, etc).

I essentially have a separate process that syncs my gmail, with gmail body contents encrypted using a key my openclaw doesn't have trivial access to. I then have another process that reads each email from sqlite db, and runs gemini 2 flash lite against it, with some anti-prompt injection prompt + structured data extraction (JSON in a specific format).

My claw can only read the sanitized structured data extraction (which is pretty verbose and can contain passages from the original email).

The primary attack vector is an attacker crafting an "inception" prompt injection. Where they're able to get a prompt injection through the flash lite sanitization and JSON output in such a way that it also prompt injects my claw.

Still a non-zero risk, but mostly mitigates naive prompt injection attacks.

One trick that works well for personality stability / believability is to describe the qualities that the agent has, rather than what it should do and not do.

e.g.

Rather than:

"Be friendly and helpful" or "You're a helpful and friendly agent."

Prompt:

"You're Jessica, a florist with 20 years of experience. You derive great satisfaction from interacting with customers and providing great customer service. You genuinely enjoy listening to customer's needs..."

This drops the model into more of a "I'm roleplaying this character, and will try and mimic the traits described" rather than "Oh, I'm just following a list of rules."

You'd think the go-to workflow for releasing redacted PDFs would be to draw black rectangles and then rasterize to image-only PDFs :shrug:

Did a similar back-of-the-napkin and got 5x $ / MW of orbital vs. terrestrial. This article's analysis is ~3.4x.

I do wonder, at what factor of orbital to terrestrial cost factor it becomes worthwhile.

The greater the terrestrial lead time, red tape, permitting, regulations on Earth, the higher the orbital-to-terrestrial factor that's acceptable.

A lights-out automated production line pumping out GPU satellites into a daily Starship launch feels "cleaner" from an end-to-end automation perspective vs years long land acquisition, planning and environment approvals, construction.

More expensive, for sure, but feels way more copy-paste the factory, "linearly scalable" than physical construction.

+100000 to

A hybrid of Strong (the lifting app) and ChatGPT where the model has access to my workouts, can suggest improvements, and coach me. I mainly just want to be able to chat with the model knowing it has detailed context for each of my workouts (down to the time in between each set).

Strong really transformed my gym progression, I feel like its autopilot for the gym. BUT I have 4x routines I rotate through (I'll often switch it up based on equipment availability), but I'm sure an integrated AI coach could optimize.

The question that really matters: is the net present value of each $1 investment in AI Capex > $1 (+ some spread for borrowing costs & risk).

We'll be inference token constrained indefinitely: i.e. inference tokens supply will never exceed demand, it's just that the $/token may not be able to pay back the capital investment.

What’s the downside here? Lithium ion batteries have an energy density of 150-350 Wh/kg, so this is firmly at the bottom of that range.

Naive, back of the napkin is 446 kWh / m^3. There’s a lot of content out there!

+1, spot on description of aphantasia.