Sign in

mike suarez

@msuarz.bsky.social
82 followers 104 following 549 posts

extreme vibe coder

PostsRepliesMedia
mike suarez @msuarz.bsky.social · 08/10/2026
yeah dear anthropic ur my fav but still ur 200 plan advertised as 4x is a f^%$ lie!!
000
mike suarez @msuarz.bsky.social · 08/10/2026
giving it js is a like giving it a loaded gun i guess lol
000
mike suarez @msuarz.bsky.social · 08/10/2026
it also shuts up if u ask it … otherwise is a perennial one more thing … i call it dj khaled
000
mike suarez @msuarz.bsky.social · 08/10/2026
jules verne wants his epoch back
000
mike suarez @msuarz.bsky.social · 08/10/2026
yeah lets make ruby more python by throwing exceptions instead of null .. xor lets try returning undefined instead and make it js :( ... man so much time wasted in dumb stuff everywhere under the pretense of the right thing
010
mike suarez @msuarz.bsky.social · 06/10/2026
except omarchy that startup produces the best non ai posts … except in italian lmao
010
mike suarez @msuarz.bsky.social · 04/10/2026
i guess u read steve yegge’s ramble round that … frankly its anthropics inception in the model … cos in real life the law system is in shambles
000
mike suarez @msuarz.bsky.social · 02/10/2026
my claude code just looks uniq lol
000
mike suarez @msuarz.bsky.social · 02/10/2026
im done being discriminated by wispr et all …
000
Reposted by mike suarez
Ruby Central @rubycentral.org · 01/10/2026
RubyConf 2027 has a city and a date. We know. You don't... Yet.🥸 Hint drops start next week, so sharpen your pencils (and your irb). Wild guesses welcome below.
013
Reposted by mike suarez
Armin Ronacher @mitsuhiko.at · 01/10/2026
We released Pi 1.0! earendil.com/posts/pi-1-0/
earendil.com
Pi 1.0 | Earendil
Today we are shipping Pi 1.0, a hardened, minimal, extensible agent harness, alongside Pi Durable, a new experimental substrate for long-running agentic applications.
1835038
mike suarez @msuarz.bsky.social · 29/09/2026
i think i would “always” want a spec, test, verification before hand even in poc spike mode … im personally allergic to the debugger … im often called non-technical by ppl that have high debugger skills
000
mike suarez @msuarz.bsky.social · 26/09/2026
stop me im going slightly mad ... the migos ...
000
mike suarez @msuarz.bsky.social · 25/09/2026
omg ... this is me talking to my greet agent live inside @freeq.at running on elita with haiku ... its so much more than a chat!!! hello world!!!
012
mike suarez @msuarz.bsky.social · 24/09/2026
u gonna keep promoting dhh ... for all the ai slop u complain this can only b explained by ai generated tweet via grok ... the uproar he's causing has nothing todo with coding nowadays
030
mike suarez @msuarz.bsky.social · 24/09/2026
woah claude finally shut's up if u ask it to !!! my stress level and anxiety are starting to go down
000
Reposted by mike suarez
Chad Fowler @chadfowler.com · 22/09/2026
@freeq.at is serious business 😂
122
mike suarez @msuarz.bsky.social · 22/09/2026
hey i have a long list of very serious tests limadelic.github.io/elita/37/rep... eg:
010
mike suarez @msuarz.bsky.social · 22/09/2026
here first test green lol github.com/limadelic/el...
000
mike suarez @msuarz.bsky.social · 22/09/2026
running the tests in freeq this was the xunit original github.com/limadelic/el...
000
mike suarez @msuarz.bsky.social · 22/09/2026
1080p looking better already lol
000
mike suarez @msuarz.bsky.social · 22/09/2026
i have these xunits github.com/limadelic/el... they run with mix test ... im figuring out how to run those interactions in freeq instead ... so freeq is like the feedback of the test running lol
200
mike suarez @msuarz.bsky.social · 22/09/2026
@chadfowler.com im doing some damage in freeq .. im planing on use it as test runner first ... so my agent tests would play as convo there in a test room lol ... im no getting the pixelation thou why 320x180 ... im switching it to 1080p
210
Reposted by mike suarez
Kwak @jeff-kwak.bsky.social · 21/09/2026
Hey folks! Finally released my game. It's a fun little one-button game. Inspired by Paku-Paku. You can play it in your browser on Itch. Check it out if you can: long-ago-studios.itch.io/pirate-moku
long-ago-studios.itch.io
Pirate Moku by Long Ago Studios
A one-button game. You're a pirate. Collect the coins. Run from the monsters. Get the sword. Eat fruit.
111
mike suarez @msuarz.bsky.social · 22/09/2026
very nice jeffy!!! ❤️
100
mike suarez @msuarz.bsky.social · 21/09/2026
"My current software factory, the Wheelhouse, is not for navel-gazing: I built it specifically to work on my game, Wyvern." ... somebody explain yegge.ai/essays/fence... what a navel is pls lmao ... he's been gazing at it for 30 yrs
yegge.ai
000
mike suarez @msuarz.bsky.social · 19/09/2026
ty i've shelved it twice waiting for better models ... i did like a full month of fable and it didnt get too far :( ... im bout to start a 3rd attempt some encouragement from u might fuel it lol ty man
000
mike suarez @msuarz.bsky.social · 19/09/2026
ive been trying to vibe an agentic platform based on that actor model elixir idea … github.com/limadelic/el...
github.com
GitHub - limadelic/elita: Agentic platform using OTP patterns for reliable agent behavior
Agentic platform using OTP patterns for reliable agent behavior - limadelic/elita
242
mike suarez @msuarz.bsky.social · 18/09/2026
dude failed fascist turing and passed when tested by fascist ai … canonical nazi
010
Reposted by mike suarez
david celis @davidcel.is · 17/09/2026
and i mean, let's be real: it's fucking hilarious that Anthropic would have even pledged this in the first place after DHH posted this shit
@dhh on Twitter: "This is why we need competition and open weights in AI. Imagine a world where only Anthropic sat as the moral arbiter of acceptable speech. Fucking ridiculous. (Grok of course did it no problem, same too with Kimi K)."

A screenshot of Claude refusing to translate DHH's blog post follows: "I'm going to pass on this one. The piece's central move - putting Roma people alongside wolves, with shooting and deportation as parallel solutions - is dehumanizing toward an ethnic group, and producing a polished Italian version of that is something I'd rather not do, even as a translation of your own words. Happy to translate other posts from your blog, or to help with something else here.
2834
mike suarez @msuarz.bsky.social · 10/09/2026
i got some simple lint rules … max 5 lines per method, 100 per module, cyclomatic complexity = 0 … i struggle to get baby steps … but i do get clean code
000
mike suarez @msuarz.bsky.social · 05/09/2026
wait what … spotify is skinnable?
110
mike suarez @msuarz.bsky.social · 03/09/2026
started using claude in a diff way with agent view ... now im having singleton agents called silos specializing in a single activity ... continuously compacting session and improving instructions on usage ... they can also talk to each other
000
mike suarez @msuarz.bsky.social · 01/09/2026
1password 1race nothing wrong with that lol
010
mike suarez @msuarz.bsky.social · 27/08/2026
idk im buying some of that ... i do use that feature a lot and it seems plausible
000
mike suarez @msuarz.bsky.social · 20/08/2026
"I am Ralph!" ... rotfl
000
mike suarez @msuarz.bsky.social · 20/08/2026
pls don't spoil the thrill of reinventing failure
020
Reposted by mike suarez
Oliver Andrich @oliverandrich.social.tchncs.de.ap.brid.gy · 08/08/2026
Elui looks pretty nice. Seems to be a nice choice if you want to implement a terminal application in Elixir. github.com/douglascorrea/elui #elixirlang
github.com
GitHub - douglascorrea/elui: Terminal UI library for Elixir, inspired by Ratatui. Build immediate-mode TUIs with OTP-friendly widgets, layouts, input, charts, and examples.
Terminal UI library for Elixir, inspired by Ratatui. Build immediate-mode TUIs with OTP-friendly widgets, layouts, input, charts, and examples. - douglascorrea/elui
022
mike suarez @msuarz.bsky.social · 01/08/2026
i got this mcp minion the idea is to still use cli from the agent … this spec makes i even more natural github.com/limadelic/mc...
github.com
GitHub - limadelic/mcp-minion: minimal mcp client
minimal mcp client. Contribute to limadelic/mcp-minion development by creating an account on GitHub.
010
mike suarez @msuarz.bsky.social · 28/07/2026
sounds pyramid shape … humans fighting to climb … leaders gotta lead and vias their leading ideas
000
mike suarez @msuarz.bsky.social · 23/07/2026
man they got time to spell llm
000
mike suarez @msuarz.bsky.social · 17/07/2026
u appreciate dhh thoughts, really? ... is politeness a thing applicable to racists aholes ideas? ... i appreciate ur politeness towards racist thoughts i guess, but i disagree with it
170
mike suarez @msuarz.bsky.social · 15/07/2026
@gmail rm this monstrosity already pls
000
mike suarez @msuarz.bsky.social · 10/07/2026
got the first scenarios for malko feature in elita ... got door, portal and malkovich malkovich ... the feature is bout controlling cli agents from other agents ... tests here limadelic.github.io/elita/cukes.... and this is the crown jewel lol
010
mike suarez @msuarz.bsky.social · 09/07/2026
been running for a whole week a fable marathon ... too much todo still to stop to talk ... but the tests ... they are coming along so pretty limadelic.github.io/elita/ these are completely vibe coded btw ... here a sample the first one .. birth ...
020
mike suarez @msuarz.bsky.social · 08/07/2026
big diff btween ppl wanting reality and sports betting … the gambling side … its like endorsing pedophiles with the excuse it gets the kids off the phone
010
mike suarez @msuarz.bsky.social · 07/07/2026
when the code adheres to my exoteric rules i can trust the agent followed my instructions … and again it takes me less than 1 second to review a file … any file … ahh i also have tests btw
000
mike suarez @msuarz.bsky.social · 07/07/2026
i have the most exoteric vibe coding rules … i expect 0 cyclomatic complexity, max 5 lines per method max 100 lines per module … do u know how easy i review 100 files of ai generated code … i just need to glance
100
Reposted by mike suarez
German Velasco @germsvel.com · 07/07/2026
I'm not a big fan of the big-upfront-plan approach to work with LLMs. Am I the only one? I wrote a bit about it in www.germanvelasco.com/blog/are-pla...
germanvelasco.com
Are plan files the new 10k-line PRs?
Many people's new way of working with LLMs is to start with an idea, have the LLM create a spec (review and approve that), then have the LLM generate a very detailed plan. At that point, people have t...
121
Reposted by mike suarez
Avdi Grimm @avdi.codes · 23/06/2026
I wrote a post about an emerging divide I've been noticing when I talk about working with agents avdi.codes/is-successful-agentic-co…
avdi.codes
Is Successful Agentic Coding a Delusion?
Somewhere in the mid-aughts eXtreme Programming (XP) practices became widely-discussed in software. From that time forward, a strange chasm emerged in the disco
252