haxel

Consider Tiny Tool Town? "A place for stupid-delightful tools made with love."

https://www.tinytooltown.com/

I also ride bike, and used to bonk, but it didn't take long for my body to adapt to using fat as fuel. Half the battle was mental: accepting that I truly didn't need to eat anything; my body already had all the fuel it needed.

Then I could wake up, ride hard for hours (hilly terrain), and feel no need to eat until the afternoon.

There's a fair chunk of irony here in that Mario is being both anti-memetic with his naming choices and contrarian in his design decisions, and yet he still finds himself dunked in the muck of popularity as the backbone of OpenClaw.

It's batteries-not-included, by design. Here's what it looks like with batteries (and note who owns this repo):

https://github.com/mitsuhiko/agent-stuff/tree/main

Perhaps benchmarks aren't the best judge.

AI studio also has a bug that continuously counts the tokens, typing or not, with 100% CPU usage.

Sometimes I wonder who is drawing more power, my laptop or the TPU cluster on the other side.

The creator is very aware. Its original name was "shitty coding agent".

https://shittycodingagent.ai/

Indeed. Currently, my primary concerns are a) surprise b) accessibility c) efficiency and d) self-containment.

Surprise, because that draws my attention better. Not interested in guardrails here.

Accessibility, because I can involve my sons without friction.

Efficiency, because using Discord lets me skip building or finding a component (for now).

And I still get a degree of self-containment because Discord is the only piece I'll need to swap out. Bonus that it doesn't have a recurring cost until then.

Yet privacy still matters to me. Despite being Discord-compromised, the detailed personal context within the prompts remains private. The data that determines the timing and topic of each message remains private as well.

I use Mistral (7B v0.2 instruct, 6-bit quantized) to generate the title-text for messages that I send to myself via a Discord bot.

Right now, I'm prompting Mistral to generate these titles in "clickbait" style. I fold the topic of the message and other context into the prompt.

My intention is to shift my attention to the message, which shifts my attention to something else I need to do, because I tend to over-focus on whatever I'm doing at the moment.

It doesn't matter whether what I'm doing at the moment is "good" or "bad". Based on probability, I should almost always switch my attention when I receive such a message because I should have switched an hour ago.

To guarantee consistent JSON output, I use a llama.cpp grammar (converted from JSON schema)

Generation is via CPU (Ryzen 5800) because it's an async background operation and also because my 1070 GPU is being used by Stable Diffusion XL Turbo to generate the image that goes along with the message.

SEEKING WORK - Mostly remote

Location: West coast Canada. Remote: Yes (occasional Vancouver onsites OK).

Frontend: SolidJS, React, Material UI, and more.

Backend: FastAPI, Flask, and more (mostly Python)

Services: PostgreSQL, OpenZFS, GlusterFS, Redis, Docker Swarm, Kubernetes, and more.

Additional: Article writing, report writing, diagramming, research, conceptual communication, and more.

I have decades of experience at all levels of the tech stack, from provisioning hardware up through services, backends, frontends, and including architecture, app design, and UIs from scratch.

When necessary, I will synthesize and implement a novel solution (while making the most of existing components).

I've been brought in multiple times as an emergency fixer and executed the rescues successfully.

I've also taken on high-level roles requiring decision-making, planning, and communication skills to go along with demonstrated high integrity.

Willing to take a crack at anything substantial.

Email: clyde6@gmail.com

Not OP, but I swapped over to Regolith a couple of months ago on my laptop and now it's on my main workstation too. Coming from Ubunutu MATE on both. Very happy with it.

It's my first time using i3 or any tiling WM for that matter. Took about a day to get used to the basics (sane defaults), and that's mostly all I've needed so far.

My one attempt to use the dedicated installer didn't work properly so instead I installed Ubuntu and then Regolith from PPA.

Developer statement

IMPORTANT: Android + Chrome only! (for now)

They say you can't manage what you don't measure.

This is a system to help people who feel fine but are worried about becoming infected by COVID-19 or any other dangerous virus. If this is you, now you can actively measure your risk and see a live dashboard in your Android phone's notification bar.

Your regularly refreshed dashboard shows your own risk pattern and compares it with your recent past as well as the risk patterns of other people. This makes it easier for you to manage both your behavior patterns and your mental health.

It's an early release by a single developer, extended from an existing platform. Please don't expect perfection.

Geographical awareness is an obvious next step. Infection/symptom awareness is another.

Feedback welcome!

I'm building a way to roughly (and simply) track your own viral risk in real-time and compare it with others.

https://howfunctional.com/viral-risk

Android-only right now. Precision is not the intention. Getting the ball rolling is.

The tracker is currently in prototype form so it may look or act a bit odd, not least because I quickly extended an existing platform.

Submission statement:

Android-only right now.

This is an experiment in roughly measuring your own viral risk as simply as possible, then comparing that with other people. Precision is not the intention. Getting the ball rolling is.

The tracker is currently in prototype form so it may look or act a bit odd, not least because I quickly extended an existing platform.

Proof of AGI kernel is on its way. Promise.

Thanks for the heads up. There's literally no ads or trackers and everything served from my domain.

I haven't tested in Firefox Focus though (haven't even heard of it), and I'm wondering if it might be because of overly tight security headers.

The other half of #1 is keep going. You have to do it every day. You have to do it for a long time. As long as you keep going, you'll naturally grow. Your mind won't stand for doing the same thing every day.

If anyone needs help getting started or keeping going, that's exactly what my project Functional helps with:

https://howfunctional.com/try/read-X-for-10-min

SEEKING WORK | Full-Stack Developer & Architect | Remote

Location: Coastal BC, Canada (Pacific time zone).

Technologies: Python, JavaScript (ES6+), Linux, ReactJS, Postgres, and Redis lately. Many others in the past.

My long-term side project: https://howfunctional.com

Email: steve@haxel.ca

I've been doing web development and backend services for 20 years now, with an aptitude for transforming projects from abstract concepts into reality. Capable of stepping in to help in a wide variety of situations, including rescue scenarios. Experience with licensed gaming systems and certification via GLI.

Available for short-term hourly or longer-term engagements.

I've used Ubuntu daily for over a decade now. My two boys use it too. Right now I'm using the MATE variant. It just works.

No issues with privacy for me because I simply uninstall the relevant packages when the time comes, if it comes.

Highly recommended!

SEEKING WORK | Full-Stack Developer & Architect | Remote

Location: Coastal BC, Canada (Pacific time zone).

Technologies: Python, JavaScript (ES6+), Linux, ReactJS, Postgres, and Redis lately. Many others in the past.

My long-term side project: https://functional.benome.ca

Email: steve@haxel.ca

I've been doing web development and backend services for 20 years now, with an aptitude for transforming projects from abstract concepts into reality. Capable of stepping in to help in a wide variety of situations, including rescue scenarios. Experience with licensed gaming systems and certification via GLI.

Available for short-term hourly or longer-term engagements.

Right now, unfortunately yes. I'll be sure to mention that limitation in the future. Supporting regions outside of Canada/US is too much of an overhead for me right now.

While the system does support non-SMS communication channels, it looks like the regulatory overhead (GDPRS, etc) means I'd need a fair bit of funding before expanding to Europe.

The website is indeed not ready for public consumption, so if there's anything I can clarify for you, I'd be happy to do so.

Edit: does this help?

The basic idea is that to solve a problem, you must take the right kind of action more often. Yet what compels us to take these actions when we are not self-motivated? Functional does, by taking it social.

1. Identify your problem or interest

2. Join or create a group whose topic is that problem or interest. The group may provide seed activities. Each activity has a target frequency (e.g. once per week)

3. Optional: add more activities you'd like to do more often

4. Tell the Functional when you do those activities

5. Everyone in your group gets to see how you're doing. Nobody gets to see what you're doing.

6. If you fall behind on an activity, you'll receive a message to do that activity from someone in your group.

The system is designed to be persuasive so that those simple messages (and other stuff) are as compelling as possible.

Take a peek at Functional, https://functional.benome.ca, a side project of mine that's in beta-testing that aims directly at your problem. Because it's also my problem, as a remote worker, and it definitely helps keep me on track.

If you're interested, I'll create a "remote-worker" group for us so we can help motivate each other from a distance.

Any other remote workers are welcome to join too, email me at steve at haxel.ca and I'll hook you up.

A benefit, rather than an incentive, would be that instead of being advertised things they don't want and when they don't want them (the vast majority of the time), they could be advertised a deal on peanut butter almost exactly when they need peanut butter, or a bike tuneup a day or two before their next bike ride, and so on.

There's a positive framing of the idea of mining (or collecting) information about yourself and selling it to your boss: it can help get you raises and promotions.

After all, when you ask for a raise you're in a wonderful negotiating position when your value is clearly quantified. When you mine/collect the information yourself, you get to choose how and when you present it. Thus you are selling the information to your boss and reaping the rewards yourself.

This seems to underscore the need for criticisms to seem credible in order to be accepted.

For example, if I write an article and I get a drive-by critical comment that offers no substantial detail drawn from the article itself, I have no idea whether or not the critique is based on a misunderstanding of what I wrote. I'm likely to dismiss the critique because it lacks credibility. I achieve nothing by reacting to a random person's hallucination of what I wrote.

But when there's a clear indication of understanding (or misunderstanding) in the critique, which to me also indicates respect, I'm happy to receive the criticism, respond to it, and perhaps adapt to it. Because then it's a real thing from a real person who cares at least a little bit about both the ideas I've expressed and me as a person.

Thanks, I appreciate your kind words. Perhaps one day I, or someone else who runs with these laws, will be credible enough for a talk like that. Right now I'm focused on developing software to let us abide by Law 1.

Good point, I appreciate your feedback.

I wonder if fixing the rest of online advertising would suffocate this technique. If, by adhering to these three laws, we gradually develop a vastly superior way to connect sellers with buyers, then destructive methods like native advertising (and I'm sure there's more) might be much less appealing to sellers.

I've been working on this problem for quite some time now, and I've distilled a basis for solving the problem of advertising - connecting sellers with buyers - into three "laws":

Law 1: Individuals must digitize their own behavior in prediction-friendly form.

Law 2: Individuals retain the exclusive right to share or sell their information.

Law 3: Sellers or middlemen who connect with buyers must compete openly for all information.

The idea is that as we adhere more and more to these laws (which we don't at all right now), we gradually solve the problems of online advertising. Many other problems too.

Blog post here: https://haxel.ca/three-laws-to-fix-online-advertising.html

There is a way to get the best of both worlds: variably-autonomous swarming vehicles. The larger the swarm, the more autonomy the car has. The smaller the swarm, the less autonomy.

Large swarms on the highway, small swarms (or solitary driving) elsewhere. Best of both!

Yeah, I would think that such a herd or swarm would be fully backward and forward compatible. Traditional cars could keep doing what they do - swarms would react intelligently to them as swarms tend to do - and there wouldn't be any need for specialized infrastructure to support pure automation.