Bilal Irfan

Software Engineering | Coding | Al
Learning, building & sharing. 💻

[Bilal Irfan @biloo0_2]

Was using /handoff by @mattpocockuk at the end of every chat, then telling the next chat to read it. Got tired of repeating it.

Built a /resume-handoff skill, it automatically finds the latest handoff and restores context.

Here's what my chat history looks like now

[Bilal Irfan @biloo0_2]

@RL_engine @thsottiaux Yeah I've already joined the wait-list

[Bilal Irfan @biloo0_2]

Gonna finish this up fast!

[Tibo @thsottiaux]

This was fixed. You know what's coming 👀

Give us 24 hours to reset the Codex rate limits across all plans.

[Bilal Irfan @biloo0_2]

@rezoundous Exactly, I think codex's models are just as good as Claude's. However its the personality and attitude that makes up for the user experience.

PS: I myself am a codex user.

[Bilal Irfan @biloo0_2]

@aditiitwt I think it’s less about the model and more about Claude's personality. It's designed to feel like a collaborator rather than a "yes-man" just taking orders, unlike other LLMs. What are you using it for, though?

[Bilal Irfan @biloo0_2]

@wahab_twts Just for context you can make up to 2 million API requests for this v4-pro model for a cost of less than $5.

[Bilal Irfan @biloo0_2]

@DanielSmidstrup It’s not a lack of capability, but a matter of principle. They want to avoid doing that in the future as well due to safety and security concerns. I remember Dario mentioning this in an interview. You know Anthropic’s whole mantra is making AI safe for humanity.

[Tyler @rezoundous]

feeling opus 4.8 or gpt 5.5 isn't good? go use gemini for 5 minutes.

[Bilal Irfan @biloo0_2]

@thsottiaux Can we expect a codex app for Linux distros?

[Bilal Irfan @biloo0_2]

No shit!

[ClaudeDevs @ClaudeDevs]

As a result of a US government directive, we are suspending access to Claude Fable 5 for all users. You can continue to use all other Claude models.

Here’s what this means for you:

Across Claude products, new sessions will run on your selected default model or Opus 4.8, and existing Fable 5 sessions will end with an error.

On the Claude Platform, requests to Fable 5 will also return an error. Please update your integrations to other Claude models.

We know this is a disruption to your workflows; we appreciate your patience and support.

[Bilal Irfan @biloo0_2]

Eagerly checking X again & again to see if @thsottiaux has posted anything about limits reset.

[Bilal Irfan @biloo0_2]

@mattpocockuk QUESTION HERE!!!
Should I use /handoff for learning purposes when already using /teach if my learning spans for multiple sessions for longer periods of time?

[Bilal Irfan @biloo0_2]

@mattpocockuk Agentic loops seem to be a major focus lately. Do you think building good loops is genuinely important, or is it being overemphasized?

[Bilal Irfan @biloo0_2]

@mattpocockuk QUESTION HERE!!
So in your video as well you mentioned you are using Opus 4.8 at medium effort. What if I use a cheaper and faster model like let say codex 5.4 at medium reasoning. Could the results be drastically different or not so much?

[Bilal Irfan @biloo0_2]

@adahstwt Only smart ones are gonna choose Sonnet 4.6 and yeah I'm the smart one.

[Bilal Irfan @biloo0_2]

@rezoundous Idc if it's a little behind anthropic's models as long as I can run more than 3 prompts on it.

Anthropic just won't.

[Bilal Irfan @biloo0_2]

@rezoundous Something to think about, however I think at this point both provide such good models that with little effort you can get similar results in both. Choice comes down to limits at this point and we all know Claude's limits can be infuriating.

[Bilal Irfan @biloo0_2]

This is to be the most relatable software & AI meme ever.

[Noah @NoahKingJr]

Anthropic with software engineers:

[Bilal Irfan @biloo0_2]

@VaibhavLLMs Switched to Linux (Ubuntu) about a year ago, never looked at windows once since.