mq
1.1K posts
mq
@mio___q
Diving into spec driven development & AI coding
Katılım Ocak 2014
1.9K Takip Edilen460 Takipçiler
English
hey surprise - you can just launch interactive in tmux and then tail the jsonl - shipped a small wrapper...ralph loop iterating to full parity rn github.com/dexhorthy/shan…
ClaudeDevs@ClaudeDevs
Starting June 15, paid Claude plans can claim a dedicated monthly credit for programmatic usage. The credit covers usage of: - Claude Agent SDK - claude -p - Claude Code GitHub Actions - Third-party apps built on the Agent SDK
English
@mattpocockuk @badlogicgames Effect is a moat.
You can Open Source all of your code without anyone being able to fork it
English
@badlogicgames Effect is an impossible product to market
It also happens to be the best way to build anything robust in TypeScript
English
English
@mattpocockuk @badlogicgames Matt, that's what draws me to errore + lintcn. Have you heard of Errore? Is it GETMO (good enough to move on)?
1. errore.org/errore-vs-effe…
2. errore.org/#:~:text=Linti…
English
AI - in theory - should make writing easier, thus expressing ideas should be a lot easier.
And yet, I don't see all that much more things that are worthwhile to read. Sure, there's a lot more junk. But I don't see more interesting eng blogs, personal tech-related blogs etc.
What is going on?
(Is this a discoverability issue? Or are people / teams not writing/sharing all that much more, indeed?)
English
@RhysSullivan Its god mode when combined with agents. But you have to be willing to go through a painful journey to get there
English
I love the spirit of this for sure! I just find CriticMarkup syntax too noisy to use regularly.
Right now I use git to track changes (not ideal) and for inline comments I butcher the md blockquote syntax like this
> @sam: my comment
> @claude: replies to me
which works surprisingly well, and only needs a couple lines of text to teach any agent. Works everywhere.
English
my advice to people making .md apps:
make it weirder!
.md is the Schelling point
what the market wants is more unique and diverse ways to interact with existing .md files
kepano@kepano
weird is good, there are already plenty of non-weird options for people who want something not weird — you should probably get weirder
English
English
@hackerbara @mio___q @kepano Cool, have many md apps have implemented changedown or it’s still early?
English
@kepano Love this push! Have a few weird ideas in the works.
But also the opposite seems important (in addition): are you aware of any efforts to standardize a few new primitives in the md spec we need for agentic workflows (eg inline comments)
English
mature things tend to get pressured into being less weird over time, so if you're starting from scratch your advantage is that you can be very weird
if you don't start weird you're probably trying to compete inside the local maximum of something more boring than you could be
English
i still read every line of code btw, often many times over
English
English
i've been trying to merge at least one PR a day using @mattpocockuk's improve-codebase-architecture skill, and it has turned into my favorite work each day.
English
@mattpocockuk Sorry but wtf is backpressure supposed to even signify here. Is the "stop" the feedback...? thats just annoying lol
English
Which term immediately makes more sense to you?
(definition in the post below)
English
@fortelabs It does have shit UX, but for such a small team its still impressive
English
Is it true that Obsidian only has around 1.5 million users!?
Compared to 225 million for Evernote at its peak?
So Obsidian has only reached 0.7% the size of the Evernote user base. It's a niche of a niche of a niche.
English
Devs: With AI, the real value you provide now is taste.
Also devs: Hey LLM, is my idea good?
English
How is everyone managing their agent SKILL.md files? Is it just chaos?
Global skills, repo-specific skills, keeping them in sync between machines, figuring out which ones you have installed, authoring new ones. What are we doing? Does anyone have a sane system?
English
mq retweetledi
Agentic architecture is a concurrency and resource management problem
Concurrency and resource management is unsolved at scale even for the best of developers. We’re talking the same problem space that led to kubernetes and orchestrators, except at 1000x higher frequency. Responsiveness measured in milliseconds not minutes. 10000 nodes in the system choreography, not 100. Vibecoders are not even remotely equipped to solve an architectural semantics problem that they cannot begin to even see let alone label.
English
Early gpt5.5 feedback:
- over defensive slop code gone
- faster than gpt 5.4, even on xhigh
- less verbose
- intelligent on low/medium
This model so far has written the best code I have read from any llm. It just gets what you want, previous models struggled with this a lot
English
I actually completely disagree- independent review should be separated from the context and decisions of the primary context, otherwise you just get the biases in the original trajectory
Dirk Kok · AI agents@dirkkok
The step where this matters is the independent review. When I want a second opinion on what the main agent built, I want the reviewer to start from the same context the main agent had, not from a handoff summary I wrote afterwards. The summary is the first place where quality degrades.
English