i dont disagree. but there is a difference between great design and entirely acceptable and shippable design.

you can always fire it up yourself and see what its all about. in my experience it generates a lot of code very quickly, that code is probably only ever supposed to be LLM maintained, not by people.

I don't think the op meant Gas Town itself (if they did, my bad), but what has Yegge done with Gas Town? By now it should have released some amazing thing if Gas Town increases productivity so much.

What has Yegge done with Gas Town? Well for one, he has posted a bunch of blog content about it which has generated chatter like this and increased his geek mindshare.

Just because he's operating in the realm of smart nerds doesn't mean he is immune to the value-inverting effects of social media.

Don't forget: also using his name and the project to pump some speculative crypto nonsense of which he was a beneficiary.

How anybody takes him (or the Ralph Wiggum guy who did the same) seriously after this is beyond me. These people should be exiled

He's now tried to create a board of a bunch of Gas Towns called "The Wasteland": https://wasteland.gastownhall.ai/

(btw I've been wondering the same thing as you and am not sure if there's another answer besides that he and people following his projects keep building projects on their projects: Beads, to Gas Town, to Wasteland, etc.)

>doesn't mean he is immune to the value-inverting effects of social media.

Or those of hype, e.g. AI hype.

I think the main thing he's produced using Gas Town is Gas Town itself.

> you can always fire it up yourself and see what its all about

I imagine it doesn't run very cheaply.

Name checks out

> that code is probably only ever supposed to be LLM maintained, not by people.

But LLMs are trying to mimic people. So if confusion is the human response, what's to stop the llm from acting confused?

A mechanical ability to look at the code without having a judgement.

oh wow you were not kidding

in general these types of attacks are still difficult to solve, because there are a lot of different ways they can be formulated. llm based security is still and unknown, but mostly i have seen people using intermediary steps to parse question intent and return canned responses if the question seems outside the intended modality.

kinda reads like an advertisement for dolt?

Dolt is definitely the most interesting thing mentioned to me.

dolt is cool until you hit merge conflicts, need non-simple migrations, or you have a process that ETLs data before putting it in dolt in bulk (huge changesets because one column changes in every row)

separate company, not aware of any association, dolt has been around for a while

i actually agree with this take; i dont see the problem that smart glasses solve. what, my phone screen isnt literally in front of my eyeballs 24/7? i have a need to be absolutely plugged into scrolling social media and consuming content so much that i just have to have the screen in my glasses? this feels much more like what tech companies want people to want rather than what people want.

Not to mention the input methods just suck major ass. They're extremely slow, error prone, and annoying. Hands are better.

And that's why I don't talk to Siri to drive my car.

I wouldn't be surprised if secured smart glasses were a useful tool in a corporate environment. By secured I mean the software stack fully controlled by corporate IT and only for use on premise. Most places will already have pervasive surveillance cameras and in a work context they might actually prove useful if used in conjunction with other computing devices.

Or maybe not. Tablets are impressively portable and the screen is probably good enough.

first let me say i agree its a solution looking for a problem

you can still take the glasses off. i dont own glasses but do use vr and the shift between putting on/taking off a headset feels more intentional than the glance at a phone. feels less addictive to me. maybe lightweight glasses and dark patterns will "fix" that eventually

You don't want your hands free?

to do what? We've already had this experiment in the form of phone calling and texting. And that's not technological because both are mature. People vastly prefer the latter. It's discrete, faster and asynchronous. In the same vein, does anyone actually use their Alexa?

To do work with your hands.

I was just in a datacenter deploying a bunch of infrastructure while coordinating with remote network operations and sysadmin teams. It was damn annoying having to constantly check my phone for new slack messages, or deal with Siri reading back messages in it's incompetent manner. I missed quite a few time sensitive messages like "move that fiber from port A to port B" due to noise or getting busy with another task and kept folks waiting for longer than needed.

In limited circumstances having a wearable "HUD" interface would be quite nice. Especially if it had great screen quality and I could do things like see a port mapping/network diagram/blueprints/whatever while doing the actual work. Would save considerable time vs. having to look down at a laptop or phone screen and lose my place in the physical wire loom or whatnot. Having an integrated crash cart (e.g. via wireless dongles) would be even more exciting.

That's just one recent task that comes to mind.

There are plenty of real world hands-on jobs where this would be quite helpful. So long as it's not connected to meta or the cloud or anything other than a local device or work network.

For a more general use-case I have what amounts to minor facial blindness/forgetfulness of names. I need to study your face for a long time over many interactions to actually remember you. Something as simple as wearing glasses vs. not can mean I will not recognize someone I've spent months interacting with multiple times a week.

I've long wished I had some way to implant something in my brain that would give the equivalent of video game name avatars superimposed over someone's head. For totally non-nefarious reasons, just names of folks I previously have met pulled from my contacts list. Obviously this is unlikely to ever be a socially acceptable thing due to recording and other potential abuses - but I have thought this for at least 25 years now - before the privacy concerns became obvious. Wishful thinking, but I can imagine myriad of uses for such technology if it didn't enable such a wide-spread number of potential abuses.

Wasn't the point of smart watches to have something even more readily accessible than a phone? I'd never want one of those dorky things, but they sell

While that may have been the original motivator, they have largely settled into a niche as a sort of fitness sensor. People do not typically use apps on them.

put something out there wade. whats your preference?

what would you prefer? i liked rust a lot as i found the compiler feedback loop pretty great, but the language was much more verbose and i found the simplicity of Go to be great, and the typing system is good enough for almost everything.

I have a feeling F# would work great, but unfortunately we don't use it at work so I can't experiment with the fancy expensive models. Only problem might be amount of training data.

Elixir works pretty well with the LLMs

i think they dont have a great moat in their individual offerings, but across square, cash app, and afterpay they offer a pretty good suite of products for the entire transaction stack.