Kishore Neelamegam (@indykish) profile picture

https://teapawt.net/t.co/8MkjXd12ev work: https://teapawt.net/t.co/EG7nsbgFUt

Chennai
Joined August 2009
Photo by Kishore Neelamegam (@indykish)Photo by Kishore Neelamegam (@indykish)Photo by Kishore Neelamegam (@indykish)Photo by Kishore Neelamegam (@indykish)Photo by Kishore Neelamegam (@indykish)Photo by Kishore Neelamegam (@indykish)Photo by Kishore Neelamegam (@indykish)Photo by Kishore Neelamegam (@indykish)Photo by Kishore Neelamegam (@indykish)Photo by Kishore Neelamegam (@indykish)
With claude, codex, opencode - for quite some time i did: context-save, context-restore now i don't.
1
51
I feel Nvidia was like Microsoft of 1990's unshaken goliath or much more. I feel this is a consipiracy to dethorne. I am sure AMD, TPUs, Maia, Titanium camp will gain.
171
Kishore Neelamegam retweeted
this is the exact issue. only a founder will take risks like this and absorb the pain it comes with they know that even they blow up, they've started from nothing once and they can do it again this is why companies die when the founder leaves
63
325
19
6,508
166,776
Well i think an interesting experience for the software development - liquid software.
I'm now building an "improvements" feature into all of my personal vibecoded applications: a freeform text input that's hooked up to an agent that yolo implements whatever is submitted. It's an interesting experience -- the activation energy barrier to sanding any rough edge or implementing any new feature becomes extremely low. Someone else who uses one of the applications I created has already implemented 150 improvements over the course of a few days. This pattern feels like a new kind of tool -- liquid software? -- and every application that doesn't have this feature now has a slightly rigid and anachronistic flavor. Why can't I just tell it how I want it to be different? Now, it's not obvious to me how this pattern can or should scale. Does it become unmaintainable after a while? (Maybe model intelligence improvement outstrips my feature appetite such that it's fine?) How does collaborative software work in this model? Can software designed this way be secure and reliable? Whatever the larger questions about broader deployment, it definitely works great in a personal context. It reminds me a little of the visions of Lisp/Lisp machines, Smalltalk, and HyperCard -- systems that the user can take control of and modify. Those visions never quite worked, though: the vast majority of software today is not end-user modifiable in any useful way. So, while this "liquid" affordance is a very small touch, the way it changes one's relation to the software feels a bit transgressive and empowering in a Lickliderian and Engelbartian way.
1
48
Codex and Dot are pretty good with gstack/egm/ orly
19
I have almost moved from claude code to codex app - cloud sessions Pain points - rust builds max out space in Mac. - manage sessions Cloud session feedback @ChatGPT - vanilla PR works - users have their cli tools / process eg: op, bun, cargo, fly, gcx, skills - gstack, egm, security(cf), orly) - save draft, publish took at least couple of hours to perfect it. - publish doesn't always show the `Environment published green tick flap` why? - unable to group my cloud sessions to a project
4
3
141
The stickyness in agent harness is not cli tooling (codex or claude code) or model. But - seamless cloud onboarding / sessions - ability to spin fast cloud agents We are may be scratching that surface now.
3
3
103
gstack bloat #gstack @garrytan - too many images that i don't need. - will ask my agent to turn off cso review.
28
1Password for agents, i see it goes to web. Lets say my use case is have codex verify: Login to planetscale db(pscale cli) See if the DB is migrated? Today: - 1Password service account in env - pscale or psql url cli access db / verify op://<vault>/item/<field> My password is leaked to claude , how does the 1Password for agent prevent this?
1
2
218
I feel the @ChatGPT @thsottiaux resets are a great way on a long term. - add new developers - upsell subs (like me, downgraded claude today, moved codex to $200)
35
Is there a speed of token response in codex in 200$ or 500$ plan? I see 6.1 Sol Max taking for ever to solve my spec.
38
neat, Converted agentsfleet to tsrs. Codex said tsrs emits the same 156 JavaScript files byte for byte, and both emitted command-line builds run as agentsfleet version 0.64.0.
Introducing tsrs: a Rust port of the TypeScript compiler 🦀 We are not forking TypeScript! ◆ In close collaborating with the TypeScript team ◆ Up to 8x faster than TS 7.0 & ~70% less memory ◆ Runs in WASM, so it works in your browser & on Workers ◆ Soon part of the Oxc toolchain npx tsrs blog.cloudflare.com/typescri…
35
Kishore Neelamegam retweeted
Introducing tsrs: a Rust port of the TypeScript compiler 🦀 We are not forking TypeScript! ◆ In close collaborating with the TypeScript team ◆ Up to 8x faster than TS 7.0 & ~70% less memory ◆ Runs in WASM, so it works in your browser & on Workers ◆ Soon part of the Oxc toolchain npx tsrs blog.cloudflare.com/typescri…
148
541
224
5,087
671,125
Two agents parallel on the same worktree - Claude Code - Codex Going to bed, both talking on <worktree>/.tmp/agent-mailbox.md
4
2
94