[HYBRID]

Eylo (a take on Hello) a voice first agent platform, multi-tenant, focused on B2B2C use-cases - https://github.com/DigiCred-OSS/eylo-os

It’s a distilled fork of a closed source product that we are offering to our customers.

I have completely moved away from Claude to Codex. OpenAI models are better at following instructions and skill use.

Claude generates more code, tries to build things that are not planned or needed, repeats same mistakes over and over.

Chatgpt on the other hand just does enough, within a project would not repeat the same mistakes, tries to guess your workflow, so you don’t have to ask it to run the same.

If you need more control over your code/projects, if you know what you want to get done, use Codex.

If you have no idea what you’re doing and are happy to let LLM drive the thing, use Claude.

Codex/Chagpt is for the competent.

"tries to build things that are not planned or needed" I literally never experience this when using plan mode upfront

I think the conclusion is a little too strong but mostly agree, Codex feels like a chisel and Claude feels like a wrecking ball.

Your words are definitely a better way to put it

One more observation - Claude is better at delegation than Codex.

Chatgpt is a better planner.

Have a discussion in Chatgpt, have Claude to plan work chunks and have GLM to deliver, Codex to review and fix, delivers an overall a better version.

Problem - It’s just too much of a context switching.

Solution - I am thinking about a new product, a collaborative workspace where I can run this workflow.

lol i do this too, but i just use my own orchestrator. recommend building one yourself, think its better that way. there's quite a few on github,

a systems language that reads like python is great!

python superset would have been perfect.

I do hope the language thrives and gets a community.

All the best to the modular team.

SaaS vendors that manage PII data needs to go through the - security questionnaires, privacy assessments, vendor risk reviews, procurement due diligence, evidence requests - sent by the customer security team.

At digicred, we were managing these in google drive and jira, but with 500+ enterprise customers, and their annual vendor reviews, this practice became unmanagable.

CAOS is a dedicated workspace that helps us manage these assessments like projects, additionally, CAOS turns the questionnaires into a system of record. Questionnaires become structured, answerable work. Finished answers and your compliance documents become a searchable, citable corpus. The next questionnaire starts from what you already said, with every claim traceable to the document or prior answer it came from.

It is self-hosted. Your policies, your answers, and your customers' questionnaires stay on your infrastructure.

Recently open-sourced - https://github.com/DigiCred-OSS/caos-os

Follow up - What have you been curious about lately?

- Canvas for Agents

- LLMs for learning.

| Idea: Don't use LLMs for summarisation. | Proof of concept - https://www.youtube.com/watch?v=1oF9M5hAazw | Repo - https://github.com/veris-pr/bwaiz

A Web-based, self-hosted database client and editor for teams - https://github.com/p-raj/collab-sqlc

Supports - Postgres - DynamoDB - Clickhouse - Redis

Primary idea is to evolve from SQL client to a Database Client, where users would be able to host queries, share queries and the work remains auditable.

Previously it was an SQL client, a PopSQL alternative. But I am trying to re-work the architecture so that it can support more databases, and services (query-as-service, query-as-reporting-job, etc).

- Web-based, self-hosted database client (a PopSQL alternative) - https://github.com/p-raj/collab-sqlc

- Loaders/Spinners should be fun - https://github.com/p-raj/pixi-loader

- self-hosted POPSQL alternative - https://github.com/p-raj/collab-sqlc

- CleanMyMac alternative based on opensource tools - https://github.com/p-raj/open-cleanmymac

- Standup meetings to comic generation - https://github.com/p-raj/standup-to-comics

- Configurable Pomodoro - https://github.com/p-raj/open-tomato

A few more closed source ones that aren't any close to be in a working state.

Do you have any examples that you've generated from standup-to-comics? I didn't see anything in the repo, and would like to see some examples before spending the time to set up myself :)

Good point, will update the repo. Though, the comics generated by the local llms all kinda suck. And, not setting this up for yourself is smart, it kind of fills up the disk pretty fast. All the recordings, local llms, etc.

I was... not prepared for your standup to comic README. Besides just being a cool idea, you gave me some things to think about and several new rabbit holes to explore.

There’s another one I’m planning to start - subtitles for your conference calls. It would generate netflix like expressive subtitles such as - “uncomfortable silence ensues”, “looks at a distance”, “playful music in the background” etc.

Internet infrastructure monitoring, monitors a third of India’s internet backbone, and a 100% of that of Bhutan. In production since 2014.

Python, C/C++, Perl, Celery, Redis, MySQL, Bare Metal.

- app - illustration

not even sure how many fields of expertise the designers have. that’s why something like roadmap.sh makes a total sense. provide someone interested a high level overview of the possible categories and then go deeper in the categories

- Sin City

- Forrest Gump