lukechilds
Luke Childs ☂️

https://lu.ke

Cofounder & CTO Umbrel.

Previously BitcoinJS, Browserify.

Public Key

npub1htnhsay5dmq3r72tukdw72pduzfdcja0yylcajuvnc2uklkhxp8qnz3qac

NIP-05 Address

_@lu.ke

NIP-57 Address

lukechilds@getalby.com

Profile Code

nprofile1qqst4emcwj2xasg3l997txh09qk7pykufwhjz0uwewxfu9wt0mtnqnspz4mhxue69uhkummnw3ezummcw3ezuer9wchszrmhwden5te0dehjuum5wghxxus2jtlra

Publishing to

nostr.oxtr.dev

no.str.cr

Author Public Key

npub1htnhsay5dmq3r72tukdw72pduzfdcja0yylcajuvnc2uklkhxp8qnz3qac

Show more details

Last Notes

2023-08-08 06:49:05

- reply

GPT-4 is very capable, the resulting bash scripts work well about 95% of the time. However there are often subtle bugs or edge cases that aren’t handled unless you explicitly tell it to look out for them.
After a few iterations of adding extra details to handle edge cases you can get something very high quality.
Results are much worse on open source LLMs but they’re catching up quickly.

2023-08-07 17:43:22

- reply

Not if you run it on your host but you can also run it in a Docker sandbox so you can just mount in directories that you want it to have access to.

2023-08-07 13:11:29

- reply

You can run it in a sandbox with Docker, also note the disclaimer.
https://void.cat/d/HSFLrikz6MdYqWmRuXVqkD.webp

2023-08-07 13:10:21

- reply

I got you fam

2023-08-07 12:43:44

humanscript is an inferpreter. A script interpreter that infers commands from natural language using AI. There is no predefined syntax, humanscripts just say what they want to happen, and when you execute them, it happens.
https://github.com/lukechilds/humanscript
This is a humanscript called tidy-screenshots. It takes an unorganised directory of screenshots and organises them into directories based on the month the screenshot was taken.
It can be executed like any other script.
https://void.cat/d/Rf3GkVQbCtg6NSfzUndkou.webp
The LLM inferpreted the humanscript into the following bash script at runtime.
https://void.cat/d/Br86EpyZev12HtsxtYka3i.webp
The code is streamed out of the LLM during inferpretation and executed line by line so execution is not blocked waiting for inference to finish. The generated code is cached on first run and will be executed instantly on subsequent runs, bypassing the need for reinferpretation.
https://void.cat/d/CCLBU6ZNWq5bXnioMGkkcB.webp
The humanscript inferpreter supports a wide range of LLM backends. It can be used with cloud hosted LLMs like OpenAI's GTP-3.5 and GPT-4 or locally running open source LLMs like Llama 2.
You can run humanscript in a sandboxed Docker environment with a single command if you want to have a play.
https://github.com/lukechilds/humanscript#install-humanscript

2023-06-24 01:50:49

- reply

It’s actually valid semver during 0.x.x releases. https://semver.org/#spec-item-4
Technically according to spec you can do anything during 0.x.x releases but it’s common convention to push the meanings across one place so:
- 0.5.3 > 0.6.0 = major bump
- 0.5.3 > 0.5.4 = minor bump
- No way to define a patch bump
This is the convention we’ve been following since initial launch of 0.1.0.

2023-05-30 20:51:41

- reply

Yes! Raspberry Pi and custom servers are still supported platforms for Umbrel.

2023-05-30 14:47:35

- reply

It idles around 5W and bursts to about 15W.
Strip our payment processor don’t support Bitcoin out of the box but we’re working on adding Bitcoin as a second payment method!

2023-04-24 06:12:58

- reply

It was a really great episode my man, thoroughly enjoyed it!

2023-02-28 19:05:29

- reply

Thanks for the tip but I’m using https://carrd.co which doesn’t support proxying. I’ll run a proxy to it on my dedicated server and add in the nostr.json file there.

2023-02-27 23:50:10

- reply

Awww man that sucks, that redirect massively simplifies my server setup. Thanks for the heads up though.
#[3] what is the security issue with redirects? Couldn’t anything bad done by a redirect also be done by a proxy?

2023-02-20 13:19:03

- reply

This is the way

2023-02-20 12:59:46

- reply

Congrats on the demotion!

2023-02-18 01:13:34

- reply

This is ridiculously cool

2023-02-17 19:58:35

- reply

Great policy, here’s one in

2023-02-17 19:43:43

- reply

#[2] is absolutely killing it

2023-02-17 19:34:07

- reply

Beautiful 😍

2023-02-17 19:25:50

- reply

Right back at you ⚡️

2023-02-17 19:24:28

- reply

I can’t zap you so have a 🤙

2023-02-17 19:19:40

- reply

Oh I can’t, you’ve gotta add a lightning address to your profile first.

2023-02-17 19:19:03

- reply

Here’s your zap you cheeky devil

2023-02-17 19:17:10

- reply

Here’s a beefy zap for your kickass work on nostream and answering all my annoying questions

2023-02-17 19:15:42

- reply

Good point, have a zap

2023-02-17 19:15:00

- reply

My bad, please accept this zap for spilling the beans

2023-02-17 19:09:46

How are you guys fighting the uncontrollable urge to excessively zap yourself into poverty?

2023-01-07 17:06:19

- reply

Yes. It’s the same as if you leak your secret key in Bitcoin or PGP. They are unusable if the secret key is no longer secret.

2023-01-07 14:33:36

Nostr as a super simple DID layer.
Interested in anyone’s thoughts.
https://github.com/nostr-protocol/nips/issues/154

2023-01-05 14:20:25

- reply

#[4] are you interested in publishing nostream to the Umbrel App Store?
A requirement is a webui but you can get around it by shipping a gotty container alongside it that just dumps the stdout logs to a web page. You can see an example of this in the Snowflake app, feel free to pinch the code from here if you’re interested: https://github.com/getumbrel/umbrel-apps/blob/master/snowflake/docker-compose.yml
Could be cool use case to run a local instance of nostream on Umbrel, connect over Tailscale in a client and get all your notes and broadcasts backed up locally.

2023-01-05 12:12:59

- reply

“Lemme quickly check a mempool” just doesn’t have the same ring to it