Jonny Olliff-Lee

@enjojoyy I know an entire team of them (20 - 30 odd), they're my colleagues and I.

I don’t know any engineer that uses Claude code anymore

@ClaudeDevs I love the unified view of background jobs, but one major issue I've found is that I can't paste into the prompt when running on Windows Terminal/PowerShell 7.x which limits it's usefulmess.

Agent view is built to run more sessions in parallel with less to keep in your head.

@niklas_wortmann A number of Skills created over the last four months or so to add some meat to the bones of this workflow: github.com/snarktank/ralph The additional skills are ones that ask questions about the plan to find anything that might be vague or ambiguous pre-PRD. Then reviews post Ralph

@DevJonny are you using some skills for that, or just handroll it yourself?

Honest question for anyone using AI coding agents daily: Do you write a full spec before prompting, or do you iterate in small steps and correct as you go? I am just not organized enough to do a proper spec-driven workflow

@quorralyne Using it daily. I create a plan/PRD with Opus, work through it, refine it until I'm happy. The hand the tasks off to Sonnet to implement. The back to Opus for review, and Sonnet to implement changes/fixes.

Anyone else using Sonnet regularly? It's my go to for daily repetitive work that doesn't need a lot of deep research like Opus.

@jpaylor @cynthiamcgillis ^^ This. Also using a GitHub Action workflow to distribute Rules and Agents into the other repos on update so all teams have the same.

@cynthiamcgillis I created a Claude plugin & plugin marketplace based on an internal GitHub repo for this. It auto updates across the team when I push a new version to the repo. It includes a hook to auto symlink rule files to the project repo on starting a session too.

How are y'all handling company-wide skills? Putting them in a repo? Does that work for Cowork and less technical teams? I feel like there has to be a better way to organize these.

#RedKite #BirdsOfPrey #WildlifePhotography #SurreyHills #NatureUK #BirdPhotography

When you're in-flight meal is just a bread roll! 📷 Nikon Z6 ii 👓 Sigma 150-600mm w/ FTZ II Adapter 🔍 600mm f/8 1/3200s ISO 500 💾 Adobe Lightroom

@svpino Get it to write a failing test first and it's simply doing 🔴 -> 🟢 -> refactor.

This works surprisingly well: • Ask Claude Code to fix a bug • Let it work and figure it out When it finishes, use this prompt: "Knowing everything you know now, delete your fix and implement a better|concise|elegant version." Claude will use the first pass to explore and learn the problem. Sometimes, this will lead to suboptimal code. The second pass will be better because Claude already knows the problem and can write a better solution.

@warnoc @clairernovotny @Aaronontheweb Installation is via Claude Code, and as the README says Claude will use the appropriate skills for the appropriate tasks. #installation" target="_blank" rel="nofollow noopener">github.com/novotnyllc/dot…

@clairernovotny @Aaronontheweb Where/how do you use this ?

Tired of coding agents not getting C#/.NET right, so I decided to teach them how. h/t to @Aaronontheweb for the inspiration github.com/novotnyllc/dot…

@Cadaren @doctorfavorito @MotorsportiveHQ

@doctorfavorito @MotorsportiveHQ Hamilton and Alonso are pretty much aging out, but where would Max go if F1 is considered the top of the mountain?

@satyajugran @iammukeshm The Microsoft one generates OpenAPI v3.0 currently (3.1 apparently coming soon).

@iammukeshm Is it implemented with Open api specs or some other standard?

If you're still using Swashbuckle in .NET 10, here's what you missed. Swashbuckle hasn't had regular updates in over a year. So, Microsoft dropped it from .NET 9+ templates and built their own. The replacement: `Microsoft.AspNetCore.OpenApi` — built-in, lighter, maintained. For a UI: Scalar > Swagger UI. It's faster, better looking, and supports dark mode out of the box. Migration takes 5 minutes: remove Swashbuckle, add `builder.Services.AddOpenApi()`, done. Read article to help you migrate -codewithmukesh.com/blog/dotnet-sw… Repost this to help a fellow developer.

@mikejulian I like the PR review functionality inside of Rider over GitHub's web interface. So that combined with being able to view the code and it's references is where the IDE fits in for me now.

For those who are 100% on Claude Code, what role does an IDE (eg, VS Code, Cursor, etc) play in your workflow, if any? Having trouble wrapping my head around this.

@buildaditya @milan_milanovic I read that as it removes the boring parts: decision making, taste, and responsibility 😅

@milan_milanovic exactly it removes the boring parts decision making, taste, and responsibility are still very human jobs

AI is here to speed us up, not to replace

@gitphysical @nicbarkeragain I've said to a few people at work that in 10 years or so I can see the rise of Artisanal Code Houses that create hand-crafted bespoke software.

@nicbarkeragain Can we have an office where nobody uses AI to code? It’s all just stack overflow and consulting other humans in the room?

"AI will generate all the code, and senior programmers will review it" is basically a mantra at this point. I'm fortunate enough to know a number of top tier programmers personally, and precisely 0% of them want to spend their time reviewing AI generated code (myself included)

@ewhauser @jessfraz I'm the same, but do the review in Rider to check other files and context. However that's now really the only reason I'm in an IDE.

@jessfraz I was always open the PR; read and comment on it in the GH web UI; and then tell the AI to fix my comments.

Here is how I do and don't use agents, idk who this will help but its worth spelling out my preferences and why: - I tell the agent to code how I would do it - If the language is one I am _very_ familiar with I feel comfortable getting it to generate very good idiomatic code that is indistinguishable from my own and doing large refactors - If the language is one I'm not comfortable with, I keep the pull request under 100-200 lines of code for the reviewers sanity since I can't discern the nuance of good/versus bad code - ALWAYS read/self review the code before opening the PR, the onus is on the AI wielder to make sure the code is up to par with what they would do themselves before inflicting their teammates - never auto open PRs because of ^ - if you do all the above you can avoid slop and not annoy your peers my name is jessie frazelle and i have not touched code in an editor since october.

@jessfraz - Exactly - Yes! - Haven't done this yet - YES!! - I do Draft PRs to take advantage of adding my own comments to be fixed - Quite! Great to see some sensible takes on Twitter!

@camsoft2000 @JohnSole651328 snarktank/ralph is a good example of this. It's ralp.sh gives you the option to use Amp or Claude as the agent, you could easily adapt to use Codex.

@JohnSole651328 Sure. These patterns are not model or harness speciric. You can write your own Ralph loop or get you agent to do it.

Having used Claude Code with Opus 4.5 a lot recently, I can tell you that for me, Codex CLI with GPT-5.2-Codex wins. It's a relief to go back to Codex; it feels like home. Not saying CC is rubbish, it's just Codex gets it done, Claue Code is too lazy, replies to my queries without re-checking code, and requires more steering and planning. Codex just doesn't need all that. I'm sure some will disagree, and it's subjective for sure, but I much prefer Codex to Claude Code. It suits the way I work.

@camsoft2000 What I love about this post is that it shows there are choices. Pick one, experiment, test it. Then try another and repeat. Use your favourite after that, or switch between.

@csinco @nickchapsas @getsome_air Oh interesting, completely missed that. When did it launch?

@DevJonny @nickchapsas That is already in motion with @getsome_air!

IDEs as we know them will become obsolete real quick