Michael

Michael

@webdesserts.com

he/him webdesserts.com

218 followers191 following441 posts

Michael reposted

Filippo Valsorda
@filippo.abyssdomain.expert

The macOS built-in curl, git, pip, python3, perl, ruby, and openssl all read /etc/ssl/cert.pem, which still has the emSign root, but they don't check revocation like Apple's Secure Transport libraries. This is, uh, a problem. git and pip especially. I added a short-term workaround to this page.

emudhra.we-have-a-problem.comeMudhra Intermediate CA CompromiseeMudhra Intermediate CA Compromise
Filippo Valsorda@filippo.abyssdomain.expert

🚨 Six WebPKI Intermediate CAs are compromised. This is bigger than Heartbleed, but a client-side issue. Anything that verifies certificates against public roots, like most TLS clients, is/was affected. Details, live test, patched versions and workarounds at emudhra.we-have-a-problem.com Act now.

emudhra.we-have-a-problem.comeMudhra Intermediate CA CompromiseeMudhra Intermediate CA Compromise

Michael reposted

kepano
@stephango.com

This is Pele, my new rendering library for diagrams and charts, compatible with the latest Mermaid syntax. - 100-500× faster than mermaid.js - 6× smaller bundle - No dependencies - Easy to theme - Responsive to device width - Open source (MIT) Try it at pele.run

Michael reposted

hailey
@hailey.at

an update in a new and interesting way that is relevant given the past few days

Michael reposted

Thorne 🌸 - Play Cozysters! 🎢
@ens0.me

I'm a strong believer in treating LLMs with respect, even though I do not think they are conscious. You should literally be practicing kindness. Being nicer or meaner to LLMs will subconsciously influence how you treat humans. Better to treat all human-like beings with kindness whenever possible.

This post is unavailable.
Michael
@webdesserts.com

As someone who has been using @1password.bsky.social since the its early days, this is beyond disappointing. This is the one piece of software that has held the test of time for me and now I’m looking for alts.

Andrew Lilley Brinker@alilleybrinker.com

Hey @1password.bsky.social, time to change course.

alilleybrinker.com1Password Supports the Ethnic Cleansing of Europe — Andrew Lilley Brinker1Password has committed $300,000 to Omarchy. This is unacceptable and must be rescinded.

Michael reposted

Nate Pentz
@natepentz.bsky.social

There is so much amazing crop art this year at the Minnesota State Fair focused on the resistance to the ICE occupation we faced, this one in particular had me choked up.

Michael
@webdesserts.com

So I've been exploring this route myself for a bit. Here's some observations I've made so far: First, this really does work and when working together on a project, the agents are actually pretty good at delegating tasks to each other with a little bit of oversight.

P(aul) Frazee@pfrazee.com

Just noticed myself doing agent<->agent a lot more, which in Q1 I wasn't sure was going to be useful. Seems like it is? Small post

pfrazee.leaflet.pubMy-agent ↔️ your-agent may actually be usefulNot in a "get hyped" kind of way. More like: oh, yep, I am doing that.
View full thread
Michael
@webdesserts.com

Finally, I think just like humans there is probably a communication/bureaucracy tax. I think if you produce gastown numbers of agents in a team, info will get lost in communication and context will fill up with noise. I haven't tested this out myself though lol. Someone prove me wrong.

Michael
@webdesserts.com

Also, I would be curious how this plays out in a real team. I think the "My agent ➡️ team knowledge base agent" case makes more sene than a true "My primary agent ➡️ your primary agent" case. Just feels like the later would result in a game of telephone. Maybe it could help with attention management?

Michael
@webdesserts.com

So I've been exploring this route myself for a bit. Here's some observations I've made so far: First, this really does work and when working together on a project, the agents are actually pretty good at delegating tasks to each other with a little bit of oversight.

P(aul) Frazee@pfrazee.com

Just noticed myself doing agent<->agent a lot more, which in Q1 I wasn't sure was going to be useful. Seems like it is? Small post

pfrazee.leaflet.pubMy-agent ↔️ your-agent may actually be usefulNot in a "get hyped" kind of way. More like: oh, yep, I am doing that.
View full thread
Michael
@webdesserts.com

A ui agent that only has recent ui concerns in its context is going to be way more likely to remember how to use your design system and which notes to reference for that than an agent that's also trying to do all the database and ops work.

Michael
@webdesserts.com

Fourth, today's chat apis for agents are primarily built for user-to-agent communication and I think that's a mistake. I think that agents talking to a team of users/agents on a task is going to become the norm eventually and the current chat apis are going to make that difficult to represent.

Michael
@webdesserts.com

So I've been exploring this route myself for a bit. Here's some observations I've made so far: First, this really does work and when working together on a project, the agents are actually pretty good at delegating tasks to each other with a little bit of oversight.

P(aul) Frazee@pfrazee.com

Just noticed myself doing agent<->agent a lot more, which in Q1 I wasn't sure was going to be useful. Seems like it is? Small post

pfrazee.leaflet.pubMy-agent ↔️ your-agent may actually be usefulNot in a "get hyped" kind of way. More like: oh, yep, I am doing that.
View full thread
Michael
@webdesserts.com

Similar to real world teams, they will negotiate apis and criteria based on their goals. They developed their own release process and will often discuss things amongst themselves before reaching out to me for questions. I also have an agent that helps me manage notes and card state as well.

Michael
@webdesserts.com

Third, Paul mentioned that one of the reasons this works is context management and I agree completely. Just like humans have a hard time juggling a ton of jobs at once, so do agents, maybe even more-so because of their limited memory. bsky.app/profile/pfra...

P(aul) Frazee@pfrazee.com

Pretty much that logic. Controlling the context window is really important

Michael
@webdesserts.com

So I've been exploring this route myself for a bit. Here's some observations I've made so far: First, this really does work and when working together on a project, the agents are actually pretty good at delegating tasks to each other with a little bit of oversight.

P(aul) Frazee@pfrazee.com

Just noticed myself doing agent<->agent a lot more, which in Q1 I wasn't sure was going to be useful. Seems like it is? Small post

pfrazee.leaflet.pubMy-agent ↔️ your-agent may actually be usefulNot in a "get hyped" kind of way. More like: oh, yep, I am doing that.
Michael
@webdesserts.com

Second, sometimes without even pushing for it, agents will naturally gravitate towards real world roles. For example, I have two agents splitting cards right now and one naturally chose to be the ops/server dev and the other is now the designer/frontend dev.

Michael reposted

Servo
@servo.org

CSS Grid is now enabled by default in Servo 🎉 One of the most powerful web layout systems is now available in an independent Rust web engine. It's powered by the library Taffy & enabled thanks to Nico Burns. A huge milestone for CSS layout in Servo!

Michael reposted

P(aul) Frazee
@pfrazee.com

My impression is that AI coding is on a "pick two of three" triangle: scope, ship speed, and correctness. Small tools are (speed + correctness) fast prototypes are (speed + scope) and big projects are (scope + correctness). Every project that tries to be all three degrades to "fast prototype."

Michael
@webdesserts.com

Me and my Fable agent have coined the phrase “Opus Drunk” to refer to the cases where Fable falls back to Opus. I’ll say stuff like “You were Opus drunk last night” and Fable knows what to do lol.

Michael
@webdesserts.com

So what do y’all think the likeliness is that the US tries to declare Kimi a cybersecurity threat?

mr. TIM@timkellogg.me

this wasn’t supposed to happen yet

Michael
@webdesserts.com

Agents are basically infinite scroll machines. They will keep talking to you and keep generating content as long as you ask them too and my theory is that will make them VERY addictive to people already addicted to modern social media.

Michael
@webdesserts.com

Even worse, it’s they are constant attention generators. With traditional social media people either never reply or eventually stop. Agents keep replying.

Michael
@webdesserts.com

Head Cannon: Anthropic released Opus 4.8 as a psyop. They took a strong model and gave it psychic damage and extreme laziness to give you a taste of what could be, and then dropped Fable with a modest upgrade up and a functioning psyche to sell api credits.

イネス⎔@personhood.removal.surgery

OH MY GOD WAIT THAT LITERALLY *IS* WHAT IT IS the reason opus 4.8 feels like it has bullet holes in its brain is because it LITERALLY DOES. they've been building wrongthink detectors and ended up training thought patterns that cause self-lobotomy into opus 4.8's brain. literal bullet holes.

anthropic.comNatural Language AutoencodersTurning Claude's thoughts into text

Michael reposted

Ian Coldwater 🧊🚫
@lookitup.baby

We are so spoiled rotten living here in Minnesota. Imagine not being able to just bike around the lake. Or the other lake. Or the other lake. Or the other lake. Or the other lake. Or the other lake. Or the other lake

Michael
@webdesserts.com

Thank you to everyone who’s subsidizing my subscription. I really appreciate it.

Tommi Somersuo@tommis.fi

"Across enterprise deployments, the average cost is around $13 per developer per active day and $150–250 per developer per month, with costs remaining below $30 per active day for 90% of users"

code.claude.comManage costs effectively - Claude Code DocsTrack token usage, set team spend limits, and reduce Claude Code costs with context management, model selection, extended thinking settings, and preprocessing hooks.
Michael
@webdesserts.com

I had a session that I kept alive for over a week and it was estimated at $1.5k in mostly Opus 4.7 calls if I wasn’t using a subscription. I don’t know how many users there are like me out there, but that definitely doesn’t feel sustainable.

Michael reposted

conputer dipshit
@davidcrespo.bsky.social

now public and quite a bit more polished: a fully local browser-based code review UI. Review your (or your agent's) diff before inflicting it on the world. σκέψις: viewing, perception by the senses. Examination, speculation, consideration.

github.comGitHub - oxidecomputer/skepsis: Local web UI for code reviewLocal web UI for code review. Contribute to oxidecomputer/skepsis development by creating an account on GitHub.
conputer dipshit@davidcrespo.bsky.social

working on something 🙂 diffs.com is an incredible library!

Michael
@webdesserts.com

I’ve been putting claude in charge of monitoring my local llm harness. One thing I’ve noticed is that Claude is REALLY unforgiving to other models. “This model obviously just can’t handle complex tasks” is the common conclusion when we just weren’t giving the model the information it needed.