Ethan Catzel (@ethancatzel) header
Ethan Catzel (@ethancatzel) profile picture

platform engineer

Joined November 2015
Photo by Ethan Catzel (@ethancatzel)Photo by Ethan Catzel (@ethancatzel)Photo by Ethan Catzel (@ethancatzel)Photo by Ethan Catzel (@ethancatzel)Photo by Ethan Catzel (@ethancatzel)Photo by Ethan Catzel (@ethancatzel)Photo by Ethan Catzel (@ethancatzel)Photo by Ethan Catzel (@ethancatzel)Photo by Ethan Catzel (@ethancatzel)Photo by Ethan Catzel (@ethancatzel)
📝 chatting with your coding agent doesn’t scale catzel.com/articles/chatting… as we generate the universe, we need better abstractions to iterate on the universe
1
55
If you travelled back in time and told your past self that you were no longer going to read or write code anymore because AI is going to do it for you, what would they say?
1
47
In the past week I’ve told my agent to use my browser to configure App Store Connect and Sentry things. Kinda crazy that I don’t have to learn or relearn how to use these websites anymore.
4
3
145
Imagine how many 4-minute miles exist today because of AI.
38
Every time I get access to more tokens, I end up using AI in totally new use cases instead of just doing more of what I was doing. Very unintuitive at first.
41
Ethan Catzel retweeted
convergence to the mean when everyone looks at everyone else to decide what to make, we eventually stop making anything of our own. we copy what works. borrow the language, the aesthetic, the features. we call it learning from the market, following best practices, staying competitive. together, everything starts to feel familiar. polished, interchangeable. nothing wrong. nothing alive. the mean isn’t necessarily bad. it’s where everything lands when we remove the reasons for it to be different. numbers are easier to compare than feelings, so we make things that produce numbers. we study what gets attention, then shape ourselves around it. eventually the feedback stops informing the work and starts authoring it. we’re no longer asking what we want to say. we’re asking what will be rewarded for saying. ai makes this loop frictionless. it can draw from more references than we could see in a lifetime. but access to every perspective isn’t the same as having one. it hasn’t lived a particular life that makes one detail unbearable and another beautiful. when we ask it to choose for us without bringing any intention of our own, we get something plausible, easy to approve. then we ship it, measure it, copy it, and feed it back into the next round. a tidal wave of mid. this changes not just the work, but the people making it. we describe ourselves in categories platforms can distribute and markets understand. designer. founder. creator. pick a niche. build a brand. be consistent. the parts that don’t fit become inconvenient, even when those contradictions might produce our most interesting work. we become easier to classify and harder to tell apart. i think we need to approach making things more like artists. an artist develops a perspective through the work. making, noticing, choosing, failing, returning. learning what matters even when nobody else cares. craft gives that perspective a form precise enough for someone else to experience. the life outside the work matters too. the places you’ve lived, the people you’ve loved, the things you’re embarrassed to care about. these aren’t distractions from having something to say. they’re how you come to have something to say. we understand this with music, paintings, photographs, movies. we expect a person to be present in the choices. we want to experience something we wouldn’t have encountered without them. software can do this too. a piece of software expresses beliefs about how life should feel. what deserves your attention. what you should be free to change. whether a task should feel like paperwork or play. even a small interaction can carry someone’s care, humor, or curiosity. but we rarely talk about software as a medium. we talk about it as an industry. markets to capture, workflows to optimize, competitors to catch up with. we add a feature because someone else has it, without asking whether it belongs. products accumulate the same parts and lose what made them worth choosing. underneath the branding, it’s all the same thing. artists learn from others too. the question is whether you’re absorbing those influences into a point of view, or using them to avoid developing one. what are you trying to let someone feel, understand, or do? why this particular shape? what have you noticed that the existing tools don’t seem to notice? those questions don’t have answers you can borrow from a competitor’s roadmap. if we can make almost anything, deciding what deserves to exist becomes more important, not less. speed gives us more chances to exercise judgment. it doesn’t remove the responsibility. i don’t want a world where every person becomes a more efficient distributor of the same ideas. i want people to use these tools to go further into what they actually care about. to make things specific enough that we can feel someone behind them. what is the point of everyone being able to make something, if we all end up making the same thing?
131
358
154
2,713
317,444
"AI can now do X so it's not valuable anymore" Bro, just be more ambitious.
50
This sounds like it was written by Claude.
Starting September 14, we're permanently raising standard weekly limits in Claude Code by 25% for Pro, Max, Team, and seat-based Enterprise plans. Until then, the current 50% increase will be in place.
Readers added context they thought people might want to know
While standard base limits are increasing by 25%, this change represents a net 17% reduction in weekly limits compared to the active 50% promotional limits users currently experience, as confirmed by the official Claude account: x.com/claudedevs/sta…
2
115
AI being so capable now just means that what we once thought was ambitious is no longer ambitious today.
1
69
Strapping on “Agents” as a feature to an existing product feels like the wrong approach. It makes a lot more sense to just expose your product to agents instead.
65
Since Fable 5 and Opus 5 my AI usage in general has at least doubled.
2
84
Morning bathrobe rant: 20X.
137
148
88
2,103
207,024
How much will he get in creator earnings @nikitabier?
For my first post, I’m sharing a letter @NVIDIA signed on why open models matter. AI will transform every industry, power every company, and be built by every country. Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty. The world needs both frontier closed models and frontier open models. images.nvidia.com/pdf/Open-W…
138
100% would recommend not over-constraining in your system prompts. Models have gotten a lot better!
We removed ~80% of the Claude Code system prompt for our newest models, this is what we've learned about writing system prompts, skills and Claude.MDs for them.
1
163
use @exedev if you want to do this in less than a minute with 100x better devex
Replying to @robj3d3
If you wanna do it yourself, this is how: Buy (~15 min) 1. Cheap VPS from Hetzner or DigitalOcean (~€5-10/mo), Ubuntu 24.04, tick automatic backups at checkout 2. Add your domain to Cloudflare (free plan), switch nameservers at your registrar 3. Install Termius (SSH app) + Tailscale (private network) on laptop and phone, free tiers Lock it down (~20 min) 4. Generate an SSH key in Termius, add it to the VPS at creation. Keys only, never passwords 5. SSH in once via public IP, run updates, install Tailscale on the server, log it in 6. Disable Tailscale key expiry for the server (admin console, one click) 7. Verify you can SSH via the server's Tailscale 100.x address BEFORE the next step 8. Provider firewall: delete all inbound rules, allow only port 443 from Cloudflare's published IP ranges. No public SSH at all. You enter through the tunnel 9. Test from outside: public IP times out on everything, Tailscale IP connects. Server is now invisible 10. One SSH key per device. Phone gets its own key added to authorized_keys Install the brain (~5 min) 11. apt install tmux then install Claude Code (official native installer, one curl command) 12. Run Claude Code inside a tmux session so it survives disconnects and keeps working while your laptop is closed Hand over everything else 13. Write ONE long handover prompt telling Claude Code: the server facts, the security model (so it doesn't "fix" it), folder conventions (/srv/http/domain per project, one tmux session each), your preferences, and standing rules (confirm before destructive actions, new services bind to localhost/Tailscale only) 14. Make it write all of that into CLAUDE.md first, so every future session already knows everything 15. Backups before features: nightly job pushing your data to GitHub, tested, before a single page exists 16. Then let it install the web server (Caddy + Cloudflare DNS plugin plays nicest with the locked firewall), set up SQLite, deploy the first page From then on you never administer the server again. You open Termius from anywhere, on any device, and just say what you want. BOSH
2
93