Dear Linux users, the wait is over! 🐧
The ChatGPT desktop app is now available in preview for Ubuntu, Debian, Fedora.
Download it and start building with Codex!
https://t.co/fgb7fs1VTs
This is the first project I started with my new skill/plugin called same-page.
It uses my Suprnova starter kit Pulsar to create a news site similar to Ground News but with my own custom algorithm. It will be an agentic managed aggregator that uses my Cognitive Prompt Architecture to create an agentic "Editorial Board". It will be interesting to see the results.
Introducing https://t.co/xcItsA9FdY with 3 free plugins and 4 paid on launch.
If you follow me, dm me your email and I will give you a free lifetime license for all paid plugins. (it will email you the license and a login with a temporary password)
I have decided to take on one or more clients. I built https://t.co/0nUgjWZ5dM in under a week including implementing Fluent and translating the website and 104 page manual to 6 additional languages.
Utilizing Pulsar or Nebula I can build pretty much any web app in Rust with React, Vue or Svelte as the frontend very rapidly. I am not limited to Rust, however. I have 30 years experience and can develop enterprise systems.
DM me if you need a web app created.
I am interested in building a portfolio. The first few clients will find that I am responsive and affordable.
Ares can pursue an open-ended objective, discover intermediate capability deficits, construct those capabilities, evaluate the resulting evidence, and dynamically create and execute new agent graphs as the problem changes, without the workflow topology being specified by the user.
@bradshannon This is great advice BUT if you are doing work for clients, they should be in the loop in regards to your other responsibilities. Don’t make the mistake of biting off more than you can chew.
Scope creep is the number one reason that your apps turn out as slop.
This is why it is important to spend time developing your idea before you start developing your app and documenting everything in detail.
Scope creep is adding features that require architectural changes to the software or data model. Coding assistants have a hard time reasoning these changes and cannot retain memory of the timeline of your changes very well. It results in conflicting instructions and causes the model to drift.
I am going to post a series over the next few days that explain how to develop a standard set of templates that you can use to start controlling the outcome rather than winging it.
The first - and most important template for you to create is your standard “best practices” document that outlines EXACTLY how you expect the model to do things. Models perform best when they have well defined and non-conflicting guidelines to follow.
Here are a list of the types of standards I am referring to and a link to a generalized template to help you get started. You can use a model to help you fill in the blanks by going through each one and printing the model to help you establish the rules of the road for each.
Here are a few examples:
- Development Philosophy
What matters the most to you when you develop a project? For me, this would be where I set rules for the road such as "Never defer features".
-- Code Organization and Architecture
- Frontend Implementation
Framework, state management, styling approach. Import ordering, component structure, naming conventions, error handling pattern. Include an annotated component example from your codebase.
- Backend Implementation
Language, framework, architectural pattern. Where validation happens, how auth works, error response format, logging conventions. Include a complete handler or service example.
- Database Schema
Database engine, access pattern, naming conventions for tables/columns/keys. Standard columns, ID format, index strategy, cascade/constraint rules. Include an example schema with related tables.
-- Development Workflow
- Version Control Practices
Branching strategy, branch naming format, commit message convention with examples. Commit granularity rules and merge requirements.
- Code Review Process
What you check for in reviews. Functions as a checklist for both humans and AI-generated code.
- Testing Strategy
Testing philosophy, unit vs. integration preference, mocking approach, coverage targets. What must be tested. Include an example test.
- CI/CD Pipeline
What triggers a deploy, what environments exist, what checks must pass, whether manual approval is required.
- Performance Standards
Response time targets, page load budgets, pagination thresholds, optimization patterns you follow.
- Documentation Requirements
What level of documentation you expect on code. Include an example showing your preferred style.
- Security Practices
Sanitization rules, auth requirements by endpoint, authorization model, fields excluded from responses, secrets management. Any compliance requirements.
---
Here is a template you can use to develop your own standards document
https://t.co/dqRUjsVY8v
👀 Seeing is just the beginning.
With Qwen-MM-Plugins, turn your favorite agent harness multimodal-native — read images, videos & documents, edit videos, work with 3D/CAD, and more.
From multimodal models → multimodal agents. 🚀
Watch it in action:
https://t.co/C2jQObXryM
Introducing Muse Glimmer, an open-weight 30B-parameter model optimized for local, always-on agent workflows.
Muse Glimmer delivers strong performance on key agentic use cases and benchmarks compared with leading models in its size category, and is designed to run entirely on consumer hardware like a Mac or PCs with performant GPUs.
In keeping with our long tradition of sharing fundamental AI research, we’re releasing model weights under a permissive Apache 2.0 license.
🧵👇