rewinfrey (@bashrw) header
rewinfrey (@bashrw) profile picture

Building @OpacityHQ (now Zoah). Previously @GitHub. MSc Software Engineering @UniofOxford. https://teapaw.org/t.co/6UVlXpFRHY

Joined May 2010
I know SF is the subject of attention in tech, but SoCal is just better, it’s not even close…
8
35
1,440
a weird inversion with LLMs is the models improve faster than the tinkerers when i see people with custom workflows and setups they're all addressing problems that don't exist anymore the person naively using vanilla codex is more likely to be experiencing state of the art
2
108
Early innings still, but I love how easy it is to use the Zoah MCP server to generate initial ideas for typography, color palettes, and branding. Then have my agent translate to a design system with components, and use those comps to build and deploy a website, all in Zoah!
3
1
8
1,595
Enjoying an amazing conference at @valiocon with my wife! It’s our first conference together!
1
13
622
Love our dev tooling at @OpacityHQ. Makes it so easy to spin worktrees up, and spin worktrees down.
1
2
352
When a review agent concludes its work, you can have it comment its findings on the associated PR in GitHub, or file follow up issues. Great for also conveying to the team and for human reviewers what was already previously caught or considered, and what is flagged as follow ups
1
54
The most commonly used agents in our tool stack are the review agents. A general review agent acts like a coordinator, spawns subagents to check on specific aspects of the diff depending on what the changes entail. But any review agent can be invoked.
1
61
We manage GitHub issues from our IDE's using custom web views that also allow us to select and trigger specialized agents from within that issue's view, or create a new worktree for that issue. Agents can run interactively or autonomously. Or can be handed off to claude harness
1
95
All internal tooling is exposed via MCP to our agents. All request and entity data is hyperlinked making it easy to nav from a single request -> actions -> events -> data from within a local web app instance. Per-request telemetry means one click to view the related data and logs
1
142
To help build @OpacityHQ, we've built a significant amount of internal tooling to help manage worktrees, run multiple local app instances without colliding, and have deeply integrated server-side and browser-side tooling to measure performance, telemetry, visualize data, etc.
1
3
3
554
Enjoying some Friday evening AI dreaming of a mix of Hiroshi Nagai and Syd Mead visual styles to produce renders for acrylic painting ideas I’d like to do by hand…
61
Congrats to the winner of the MacBook Neo, and amazingly is a designer!
Insane! @OpacityHQ had the largest turn out at all of @stripe Sessions! Thanks everyone for hanging out with us!!! Mind blowing!
1
1
7
1,636
Replying to @bashrw @_clem
Also so impressed by the overall thoughtfulness and design approach by José Valim and team, and loved his ElixirConf presentation about the various things considered for Elixir’s type system. Cool balance of technical feasibility and working with community youtu.be/giYbq4HmfGA
1
29
What an amazing week learning concurrency oriented programming via Erlang’s concurrency model from @FrancescoC! With a surprise guest presentation by @rvirding! Loved the focus on scalability and reliability, functional approach, process as core unit, and node to node msg passing
2
14
767
Replying to @defnotbeka
I’m sorry you experience this. Trying what I can… ❤️
1
92
I wasn’t expecting this old Fortune magazine from 1967 to showcase a Ford ad about using computers to aid education. The copy almost reads as though it could be written today for many edtech products.