James Powell

DUTC Open Science Scavenger Hunt Welcome to the DUTC Open Science Scavenger Hunt! Explore key platforms and resources in the open science movement. Answer each clue based on the websites provided. RECOGNITION & PRIZES 🥇 First to co...

Búsqueda del Tesoro de #CienciaAbierta
Organizada por Don't Use This Code @dontusethiscode.bsky.social
🙋‍♀️🙋‍♂️ Es una actividad educativa y divertida (en inglés) que guía a quienes participen por recursos fundamentales para la investigación abierta.

Puedes participar en este enlace 👇

As I understand it from the interface, we can delete memories, to eliminate risk it retains an unproductive memory.

That said, I am not a very advanced or very sophisticated user of these tools, but I don't really see what memory offers, other than the risk of being fooled more effectively.

What am I even doing here?

I recently became a heavy user of LLMs. I have, at points, been a daily user (though I have yet to use them when coding.)

An obvious risk is that these tools are quite convincing, even when dangerously wrong.

Perhaps an equal risk is how readily we manage convince ourselves…

In fact, I asked it to explain the output, and what I got back was almost literally a horoscope.

It appears that the corroboration mostly references the conversation itself. This is probably because I have memory disabled. That's a horoscope if I ever saw one.…

I tried this prompt and strongly suspect the output (like much of this kind of unrestrained prompting?) is a horoscope.

Should I be embarrassed by how many times `pacman -Qqn | pacman -Syyu -` has fixed things after I have made a dumb, avoidable mistake (e.g., like rebooting in the middle of a `pacman -Syyu` whole-system-upgrade…)

In single-user mode, the fix is easy, obvious, but a bit irritating.

# using Arch install media
$ cryptsetup luksOpen /dev/nvme1n1p2 root
$ mount /dev/mapper/root /mnt
$ mount /dev/nvme1n1p1 /mnt/boot
$ arch-chroot /mnt /bin/zsh

# inside chroot
$ pacman -S pam{,base,openssh,sudo,util-linux}

#linux #puzzler

How do you recover from `rm /etc/pam.d -rf` on a machine you control *without* booting into single-user mode?

Friends, I'm #OpenToWork and I'd appreciate any leads, and networking opportunities.

Don't miss this rare opportunity to work with the one and only Mariatta.

🎯 Lead Software Engineer roles
✅ 20+ years experience
✅ award winning open source contributor
✅ Expert in #Python #Django APIs automations

# hmm

from dataclasses import dataclass

@dataclass
class T:
value : int = None

obj = T()
for obj.value in range(3):
print(f'{obj = }')

What better way to start a Thursday than by announcing our #PyTexas 2025 Keynote speakers! We're overjoyed to be joined by @mariatta.ca and @kjaymiller.com this year as our Keynoters! 🎉 🎉 🎉

Want to join us? Early Bird Tickets go on sale tomorrow!

Posting to Bluesky using Python API.

Code: gist.github.com/dutc/50f54d50afa16c16ddc78c463d1fdac5

API documentation: atproto.blue
API examples: github.com/MarshalX/atproto/tree/main/examples

@marshal.dev

Security means securing people where they are https://blog.yossarian.net/2024/11/18/Security-means-securing-people-where-they-are

if __name__ == '__main__':
with open(__file__) as f:
code = f.read().replace(' ' * 4, '\t')
run(main(handle=environ['HANDLE'], password=environ['PASSWORD'], code=code))
### 04/04

for msg in messages:
post = await client.send_post(
client_utils.TextBuilder().text(msg),
reply_to=models.AppBskyFeedPost.ReplyRef(parent=last_ref, root=root_ref) if root_ref else None,
)
last_ref = models.create_strong_ref(post)
root_ref = root_ref or last_ref

### 03/04

async def main(handle, password, code):
parts = [*chunked(code, size=290)]
messages = [
f'{txt}\n### {idx:02d}/{len(parts):02d}'
for idx, txt in enumerate(parts, start=1)
]

client = AsyncClient()
profile = await client.login(handle, password)

root_ref = last_ref = None
### 02/04

from asyncio import run
from os import environ

from atproto import AsyncClient, client_utils, models

def chunked(text, *, size):
buf = []
for ln in text.splitlines():
if sum(map(len, buf)) + len(ln) >= size:
yield '\n'.join(buf)
buf = []
buf.append(ln)
yield '\n'.join(buf)

### 01/04

YouTube video by Curry On! Ron Pressler - Pull Push: Please stop polluting our imperative languages with pure concepts-Curry On

Ironically the language that has the best answer today to async/await is … Java. And that's I think mostly thanks to Ron Pressler. This video from years ago is a must-watch for anyone who is interested in the failures of async/await and rethinking that entire idea. www.youtube.com/watch?v=449j...

I made a Python core team starter pack:
go.bsky.app/AgFvebt

Let me know if I've missed anyone!

#Python #CPython #PythonCoreTeam #StarterPack