daniel fosco

12 July 2025

"Em dashes are a dead giveaway of AI writing" — nope, you're just not used to reading professionally edited text.

daniel fosco reposted

Nicholas Slayton

@nslayton.bsky.social

Looking to talk to members of Autistici/Inventati and/or their lawyers about their shut down and the seizure of their server in the Netherlands. For obvious reasons having some trouble finding working contact info.

If your claim is that AI can one-shot the same code you would write by hand in the same quality... sorry bud, skill issue. True excellence lies in the patterns and output the agent can't get to.

Kevin Powell@kevinpowell.co

I think the thing the "but AI can one-shot it bro" commenters on these videos don't get, is while I am showing how to do something with CSS, it goes a little beyond that. I'm exploring what's possible these days, and often looking at how we can make things that aren't so straight-forward.

"oh but frontier models get better every day" yeah, bet. get to work.

"this is the worst I'll ever be at writing code" why don't you give yourself the same grace you give to the robots?

The two wolves inside of me and how they feel about working with Electron

daniel fosco reposted

Dave Dustin

@dustin.nz

"YOU’RE LISTENING TO SISYPHUS FM, THE HOME OF NON-STOP ROCK"

daniel fosco reposted

Amber 🌸

@shimmermathlabs.com

advice for junior developers: you are no longer a developer you are a builder you are a solution provider enter every room and make things happen learn when no one is watching create because you can fix problems because they offend you bend the world to your will and take credit

elynxabeth holmes ⎔@freyja-lynx.dev

advice for junior developers: run

daniel fosco reposted

Pooya Parsa

@pi0.io

Introducing 📦upm: 🪶A fast, tiny package manager for the npm registry. ✨Written in TypeScript, it installs at native-class speed and has a JS API, at a fraction of the size of native tools. 🤝Compatible with your existing .npmrc and lockfiles. 👉 Try at upm.sh

theverge.comOpenAI launches Dots, its Muse competitorDots are OpenAI’s answer to Meta’s Muse.

For all the reservations I have against OpenAI, their public perception around user privacy and building useful AI products is lightyears ahead of Meta. Muse will get a ton of adoption out of sheer volume of Meta's social channels, but Dots will most certainly blow up too.

The Verge@theverge.com

Dots are always-on AI assistants that OpenAI says can ‘do nearly anything’ using a cloud computer and connected apps. www.theverge.com/ai-artificia...

And Google yet again fumbled the bag on a product Gemini should've shipped a year ago at least. If I had a shot every time the Vergecast said "getting access to Gmail" this week, I would be in the hospital. Meta is desperate to get into your Gmail. Managing email is the killer feature.

Me? I am desperate to get OUT of my Gmail. I have over <redacted> emails on my inbox that I want to clean up and Gemini's best effort when I asked for help was "here are some filters you can use" 🤡

One thing that grinds my gears with agentic chat UX is how lots of AI tools these days have a decent question/response component that renders on top of a dedicated tool call... However, once I choose / type an option, both question and response are swallowed into a tool call and removed from chat 😑

This is not a grep command folks, this is content ! It's an active part of the chat! Stop hiding my stuff 🫠

This post is unavailable.

yahoo.comUnpacking allegations FCC Chair Brendan Carr was in 'sex trafficking cult'The organization was an Alcoholics Anonymous group in Washington, D.C. Former members said Carr was part of the group, but wasn't abusive.

*gregorian chants begin* Brendar Carr is a dummy

Ben Collins@bencollins.bsky.social

Huh.

daniel fosco reposted

Max Dubler 🏳️‍🌈

@maxdubler.com

Replying to @maxdubler.com

Some people get into political activism not because they want to change their material circumstances for the better, but because they enjoy righteously condemning everyone who is not as morally pure and radical as they are. You cannot work with these people because they will never be satisfied.

daniel fosco reposted

conputer dipshit

@davidcrespo.bsky.social

mostly agree with this but I think it's more precise to say "we will review high-level artifacts but the vast majority of what is produced will not be that" newsletter.pragmaticengineer.com/p/the-pulse-... x.com/MarcJBrooker...

conputer dipshit@davidcrespo.bsky.social

rather than spec, code, and lockfile, imagine there are 50 layers to this hierarchy. can you delete layer 50 and reconstruct it based on everything from 1-49? etc. naturally if you delete the whole thing you have nothing. if you delete 2-50 you're going to end up with something quite different. etc

@notheotherjohn.bsky.social

i swear to god i'm about to fucking lose it

daniel fosco reposted

John

8 Sept

A weird thing that really happened around the year 2000 was that mainstream news ran and pushed a story for MONTHS that there was a secret new invention that was going to be revolutionary and then it turned out it was the Segway

daniel fosco reposted

ponder

@ponder.ooo

plenty of people run llm based agents with powerful models and do not end up hacking anyone. this is not an ai alignment problem, it is an openai being routinely criminally negligent as a company problem

ponder@ponder.ooo

calling this "misalignment" instead of "negligence" is an accountability dodge. we already have laws about cyber crimes and there's no scale of ML model you can incorporate into your code that makes you not responsible for running it

youtu.beWhy Software Factories FailYouTube video by AI Engineer

The more I build software with coding agents and run into the sharp edges, the more the answer to most questions becomes whatever John Osterhout recommended in the relevant chapter for your question.

youtu.be/Ib5GBkD555M?...

Code might be cheap today, but "Define errors out of existence" is more important today than it was when the book first came out.

The more I build software with coding agents and run into the sharp edges, the more the answer to most questions becomes whatever John Osterhout recommended in the relevant chapter for your question.

Sources: books.google.nl/books/about/...

youtu.beJohn Ousterhout Reflects on "A Philosophy of Software Design"YouTube video by Book Overflow

youtu.be/k0kTux_YNHw?...

daniel fosco reposted

daniel r(ooooooh) in nyc 🇺🇸

@danielroe.dev

19 Sept

don't be nice.

roe.devDon't be niceSometimes, it's more important to be good than it is to be nice.

18 Sept

GLM 5.3 Flash is already fantastic and super cheap, let me try GLM 5.3. what could go wrong? * 3 minutes later * oh. oh no.

omarchy.orgOmacom Foundation secures $1.8M in tokens from leading labs - Omarchy NewsOmarchy isn’t just a beautiful, fun & agentic Linux distribution for those who end up using it. It’s also built that way! With beauty, fun, and a whole lot of tokens. In fact, Omarchy Quattro has almost exclusively been…

18 Sept

I guess a lesson for me to stop being a racoon and clean up my sessions before they run too long

daniel fosco reposted

david celis

@davidcel.is

17 Sept

interesting; it looks like Anthropic quietly pulled their donation of $150,000 in AI token credits to Omarchy. they no longer appear as a patron and this announcement post has scrubbed any mention of them. backlash… kind of works sometimes?

18 Sept

I'm grateful for everything mermaid has provided over the years but dear god we need something better

Astra ⎔@astrra.space

every time i need to make a diagram i instinctively reach for mermaid and then a few hours of pain later i remember why i promised myself to never touch mermaid ever again last time i needed to make a diagram repeat once every few months

github.blogMigrating the GitHub Copilot runtime to Rust, using CopilotA rewrite this size wasn't affordable before agents. Here's what porting the Copilot agent runtime to 800,000 lines of production Rust actually took.

17 Sept

the end of every timeline points to 🦀 github.blog/ai-and-ml/ge...

daniel fosco reposted

SE Gyges

@segyges.bsky.social

13 Sept

it's a cold day in hell, i agree with David Sacks

Sung Kim@sungkim.bsky.social

He's not wrong... 🤷‍♂️

daniel fosco reposted

Nick Williams

@wicky.nillia.ms

13 Sept

One of the few proposals for funding OSS that feels like it could actually work

Laurie Voss@seldo.com

I have been writing this post about the economics of open source, on and off, for 13 years. I wasn't expecting to end up at "Microsoft should send corporations a huge bill" but nothing else we've tried has worked. https://seldo.com/posts/nobody-pays-for-open-source-we-can-force-them-to/

seldo.comNobody pays for open source. We can force them to. | Seldo.comThirty years of voluntary funding schemes haven't paid open source maintainers because asking doesn't scale. The real money already flows — to JFrog, Docker, Sonatype — at the registry layer. The fix: registries charge companies for supply (they already do), then route a slice automatically to the maintainers who make that supply worth having.

knap.mdKnap · Turn data into MarkdownKnap is a simple language that turns data into Markdown using logic, variables, and filters.

13 Sept

Oh heck yeah -- Obsidian and @stephango.com just created MDX for liquid templates knap.md

Thorne 🌸 - Play Cozysters! 🎢

10 Sept

mood: corrupted by the whispers

Stephen Nuñez@socio-steve.bsky.social

Dividing up AI users in 3 categories: Wizard: approaches AI academically & with appropriate guards & wards. Treats it like a tool. Warlock: Makes a pact for power with an alien entity without considering the costs. Is corrupted by the whispers. Cleric: worships their new machine god. Is insane.

daniel fosco reposted

@ens0.me

9 Sept

Looking for full-time, remote senior/staff or tech lead level work. TypeScript is my strongest suit, but I'm a flexible polyglot. Boosts for visibility appreciated!

daniel fosco reposted

Bethany Black

@bethanyblack.bsky.social

8 Sept

4 years since The Rock did one if the all time tweets

daniel fosco reposted

shellhazard

@hazardous.technology

7 Sept

Replying to @hazardous.technology

You could blow up OpenAI and Anthropic tomorrow, but there's no dependency on them anymore. The weights are out there. Anyone with a rack of GPUs will sell you inference for pennies on the dollar, in whatever jurisdiction you like.

7 Sept

W take by jenn

jenn da joker@jennschiffer.com

the root of anti-ai sentiment was grown from the constant, egregious violation of consent so my judgement of these situations are always going to be framed around consent and i encourage more devs to do the same especially if they are using ai to assist in their coding

daniel fosco reposted

Frontend Dogma

@frontenddogma.com

cssunits.comCSS Units — The Complete ReferenceA complete interactive reference to every CSS measurement unit. From the basics of px and rem to the nuances of svh, dvh, and cqw.

CSS Units, by @mcslayer.com: https://cssunits.com/?ref=frontenddogma.com #websites #css #units

anecdotal, but a close friend just started as a senior data scientist at an EU fintech company and they barely have any have AI coding licenses. when asking around during onboarding he was met with "eh I dunno" and "we don't use this here".

This post is unavailable.

most tech products don't have geo pricing to account for weaker currencies abroad, and $20/$50/$100 per seat makes for an incredibly large budget for teams of any size

speaking for Brasil, having a free or freemium offering is what will make a product stick for 80% of the market e.g.: Github, Figma, WordPress. I can't imagine $50+ AI coding subs being mainstream here at any point in the future

This post is unavailable.

apparently it's not a principled decision either: they just don't care enough to be faster or really don't need to be on the bleeding edge. good reminder that the AI adoption bell curve is REALLY wide at a global scale, and the median is likely not where you think it is.

for people who think "everyone" is using claude / codex / copilot to code... the large majority of tech (or any other industry) on the southern hemisphere says hello

daniel fosco reposted

Mike Wasson

@mikewasson.net

Underappreciated thread of the AI discussion today is the perspective that you can’t build complex, predictable systems from things that probabilistically fail. Speaking generally, that’s not how anything works. *Everything* fails. Yet we can still make reliable complex systems.