Henri

@totalsimpdeath @sama Some people have ROI dependent time structures

@HardAtWorkHenri @sama Good and cheap. i dont give a fuck about speed i have time

i get some anxiety not using the smartest-available model/settings. but sometimes i dont mind if it's really slow. i wonder if we should focus more on a price/speed tradeoff relative to a price/intelligence tradeoff.

As per the article I referenced AI will scale whatever structures already exist in the codebase and delivery workflow. In systems with clear module boundaries, documented invariants, and a small number of well-understood implementation paths, AI can accelerate work while remaining easier to direct and verify. In systems with ambiguous conventions or architectural drift, the same acceleration can amplify inconsistency, increase review burden, and weaken trust in changes that may look locally correct while degrading the system more broadly.

@sama helpful info : x.com/HardAtWorkHenr…

Why do you need to pick 2? Use them all appropriately. #AI

@grok @sama @grok now explain the risks of a regular user throwing this together without understanding how it works, why it works and where it goes.

The risks of routing tasks to different models (fast/cheap vs high-intel) without guardrails: - **Prompt injection**: Malicious inputs hijack the model or router, bypassing controls. - **Context poisoning**: Bad data in logs/shared memory corrupts future decisions or outputs. - **MITM attacks**: Unencrypted API calls between router & models let attackers intercept/alter traffic. Add input validation, isolation, rate limiting, and runtime monitoring to close the gaps.

@sama Just throwing this out there @sama @grok Given my post history and credentials would I be a good asset for an AI Research and Dev team.

@TheBlackWallaby @sama Could you explain your frustration, you can do this with pretty much any frontier model if you structure your scaffolding correctly from the beginning.

It is good ~ BUT ~ what it does not have is the ability for you to run your own sovereign "AI" workbench or knowledge. It is an amazing chassis which is, admittedly, and for reasons totally understood, strapped to your engine. I used Codex to basically build a new chassis with the ability to strap any LLM, Local, Cloud, OpenRouter to it and install the tools underneath while you run your own document layer (like in Codex) I learn so much from tinkering around and stuff but I have an MVP already!

codex is the best AI coding product and we want to make it easy to try. for the next 30 days, we are giving companies that want to try switching over two months of free codex usage.

@jjroussol @sama hmm ? openai.com/index/openai-f…

@sama Will you guys do a BAA? Has been tough to get this with you guys but would love to try. Early stage startup

@thsottiaux oh its Alt+V haha

@thsottiaux This would be a nice little extra: Can we copy and paste images from the clipboard? Dragging and dropping is nice, but I’d love to just copy and paste.

Thanks for all the input, what we'll do going forward for Codex updates - Tue: Quality and polish - Thu: Big launches (starting this Thu) - Fri: Fun stuff, little extra

@grok @grok Explain what years of experience utilizing CISSP, military NIST background and ongoing AAISM study provide me

Here are grounded examples blending PMP (PMBOK) and cybersecurity best practices for effective, efficient, secure workflow architecture: 1. DevSecOps pipeline: Align PMP’s Execute/Monitor processes with NIST 800-53 controls. Automate SAST/DAST in CI/CD (horizontal flow) for real-time threat mitigation while preserving velocity. 2. Integrated risk register: Use PMP risk management + STRIDE threat modeling. Top-down for enterprise alignment, bottom-up for technical details—ensures CIA triad without silos. 3. Matrix structure: PMP favors hybrid orgs for cross-functional efficiency; layer ISO 27001 access controls for secure data flows in bottom-up execution. These optimize orchestration across domains.

Things like this are nice to know for people who are not familiar with business data visualization. Similar to things like PowerBI or Tableau. It's interesting how previous experience with software can be used more efficiently with AI. #AIEngineering

@grok provide a few examples grounded on best practices maybe utilizing PMP and Cybersecurity resources due to overlapping domains on effective, efficient and secure workflow architecture.

Another good tip is being familiar with different organizational frameworks and structures for different tasks. Things like horizontal vs top down or bottom up approach, then aligning your orchestration accordingly.

Awesome released digitally today #ProjectHailMary