rich

I’m starting a weekly “Could I..?” series. One question. One experiment. Public notes, code, mistakes and results. Mostly AI, dev tools, games, automation and data.

@CastAsHuman x.com/i/status/20706…

@RichDevLab You think that’s a possibility?

It's been a few hours since OpenAI's announcement. How do you feel about it? Do you think we are going to wait for a long time before we get our hands on the model?

Seeing a lot of 20-user apps with Postgres, queues, and k8s. For most SaaS, SQLite in prod is fine.

I can't wait to see my timeline full of people trying to charge $19.99 for 20 lines of markdown

The tool I use most has no landing page. Just a janky internal flow that saves me hours every week. Still more useful than half the polished stuff I almost shipped.

@rlakhnyuk the part people skip is that spending on creators and ads forces clarity fast if a paid click lands and nobody buys, your offer or page is the problem, not your motivation

$0 MRR is a skill issue. That's good news because skills can be learned. But not by reading more tweets. Build things. Launch often. Post every day. Spend money on creators and ads. Make bets. You’ll probably lose money for a while. But every failed experiment builds the instincts that eventually turn you into a Badass Operator.

@thsottiaux @adonis_singh cheeky

@adonis_singh Nevermind

gpt-5.5 is 2 months old btw, they definitely have a much better model internally (likely better than fable)

how have I only just seen this cool little Codex addition? lovely bit of UI interaction

@Tech_girlll because now you're debugging the prompt and the code path at the same time this is why vibe coding doesn't replace traditional knowledge of patterns, proper implementation etc.

Nobody wants to say it but vibe coding is harder than actual coding.

great to see this finally confirmed set phasers to fast mode, reset inc

@kylegawley i've seen this a few times where the first version was early, not wrong. infra gets cheaper, habits change, then the same idea suddenly works

One of my biggest mistakes is shutting down products completely Sometimes things change over time, tech improves etc and projects can get a new lease of life

@tibo_maker this is why plain text launch posts fall flat, they tell me what shipped but not why i should care. auto-picking the most impactful thing is the smart bit

most founders are bad at announcing what they ship tbh, it's hard to turn a feature release into an event most announcements are basically: "we shipped X" so we just added motion-graphics in Revid I used it to create this video, just 1 prompt NOT A SINGLE MANUAL EDIT. what I am doing now: - connecting it to my Mission Control agents - monitoring my git repo - picking the most impactful thing we shipped - automatically create a video to announce the new feature wild

@jturntdev my issue isn't even the delay by itself, it's that 'ai for everyone' sounds a bit hollow if frontier access is basically invite only

Sooo let me get this straight. GPT 5.6 was delayed to the masses, but not the ultra elite? We’re only deserving of the dumb watered down models, whilst they access frontier? Cant just be me who only sees a negative outcome to this?? What happened to OpenAI’s “AGI for everyone”

@itsolelehmann i agree, but the tricky part is knowing what actually made it work

you really only need a handful of bangers to grow a following online you should obsess about understanding what makes a good post you can get 1000x returns from 1 good post vs 1000 average ones which is crazy to think about

@sharyph_ this is clean subagents being context-isolated is the bit most people miss, otherwise the main agent starts dragging old newsletter context into everything

How to set up Claude Skills so it runs like a full content team: 5 folders. That's the entire system. 1. CLAUDE.md → Memory. Your brand's constitution. Voice rules, audience, offer details. One global file for all projects. One local file per project. 2. skills/ → Knowledge. Reusable content workflows Claude auto-invokes by matching the task. No commands. It just knows what to write and how. 3. hooks/ → Guardrails. Rules that run before and after every output. Keep tone consistent. Block off-brand language. Auto-check CTAs. Deterministic. Not AI. 4. subagents/ → Delegation. Isolated agents with their own focus. A newsletter writer that only sees your outline. A repurposing agent with content-only context. Keeps your main workflow clean. 5. plugins/ → Distribution. Bundle everything into one install. Every project starts with the same voice, same rules, same agents. Aligned from day one. This is your Content Operating System. Five layers, one stack. To learn how to build yours: → Go to newsletter.thedigitalcreator.co/p/newsletter-c… → Subscribe free — just drop your email. → Open the welcome email and get the free resources. Repost ♻️ to help a creator in your network.

generalists are set up insanely well right now if you can think across: - tech - business - people and connect things fast, that's a goldmine

@Road_Kill11 i'd add ci too it still assumes one clean human change moving through a line. with agents you get a pile of partial edits, and the hard part becomes proving what is actually safe to merge

unpopular opinion: loads of devtools that people are heavily reliant upon are not built for the future and are ripe for disruption - git - worktrees are suboptimal ways to make parallel code changes, commits don't fully represent intent well enough or have enough first class agent primitives. stacked diffs help here but there is a lot more that can be done - github - terrible uptime, need i say more? - code reviews - mostly just rubber stamping now with AI, needs to reduce cognitive debt of reviewers to be more useful. github PRs do not help with this - IDEs - built for single threaded synchronous work. i think there will be less of this going forward anything else i'm missing?

@zats oh this is really cool

I wanted faster iterations on codex responses instead of asking questions one by one, so I added sticky notes - annotate codex response with your questions and let it answer all at once!

@zdogmode feels useful for structure, risky for voice once every post starts using the same wording and cadence people can tell even if they can't name why

has AI improved your content as a founder or made it sound less like you? genuinely curious

@kylezantos i'd take this just so people can say 'agents' and not mean 4 completely different things once you know whether someone builds this stuff or just tried it once, the whole thread gets way less wobbly

We need some kind of official AI proficiency badges on here. I’d like to stop talking over the heads of 2nd year magicians, and conversely not over-explain to legit wizards. Really would like just speak your appropriate AI dialect.