The X Handle Marketplace offered a free handle change, and I picked @NoHypeGuy.
It's still me, same account. The name just fits what I'm doing here, which is posting about AI and technology without the hype.
Allow me to explain this more in depth. I'm not talking about simply prompting to fix surface-level bugs. I'm talking about prompting with the goal of fully understanding the architecture:
- Running it so it understands all the dependencies in every direction, inside and outside the architecture.
- Writing SOPs around it and
- Formulating a plan to build a knowledge base that we could consistently check our work against and that encompasses all this knowledge of the architecture
- Bringing in outside models to also reason and fill in any gaps you might have initially missed
@ParamSiddh What if the job has an interview component or an interview-skill-adjacent component to it?
Maybe partaking in Q&A sessions Or regular reviews.
Interesting new model showing up for ChatGPTβ¦
β¦although I'm pretty sure I remember hearing about Aurora several months ago, so maybe it's a specialized version of an older model.
Uggh π£
OpenCode was hacked and API's were compromised. Good folks over there at OpenCode, but this is the reason I run with my own custom architecture.
My keys are local, behind several layers of security.
We had a security breach last night related to a large scale distillation attack that impacted 1.4% of OpenCode Console users
We have notified them all this morning via email. If you did not get an email you were not impacted, it was a subset of users who had external providers linked
The incident is ongoing so we cannot share more details at this point but we are working with Anthropic, OpenAI, and Google to remediate
@pbshgthm Interesting.
If it's happening on different devices, I'm wondering if it's something with your ISP. Trying going to a library/Internet cafe and see if it's still happening to you.
I use Claude all day (including 30s ago) and it answers me the same speed it always has been.
What env is it in?
If it has a web build, point a browser agent (Playwright plus Claude Code or Codex) at it with a test script of goals to try. Start small and then scale up.
If its engine, use Unity ML-Agents (or Godot RL Agents) to train a bot that plays the game in-engine. You can then have it report on where the bugs happen.
One key thing I forgot to call out: I can hit that coding model and flip through the different models I have. It will change my hooks and directives to send coding for the next task to that agent that's on the window.
Let me start by saying, I'm not endorsed by Elgato.
One of the most important parts of my setup is my Stream Deck.
I have it fully customized. It shows me all the models that I regularly work with, gives me the visibility I need to make quick decisions, and hotkeys the to make them.
The Menu below pops up when I click on my coding window. I have other menus for other things I click on.
Yeah, doing the same thing over here. I run 8 agents as well, and I have hit this in the past but don't anymore.
The only thing I can say is that when something happens, peel back the onion, find out the 'why's' behind it, and ask the tough questions. My favorite one to ask Claude "Open a ticket (yes I have a personal ticket system), let's figure out what happened and how to guard against this from ever happening again."
And don't take Claudes fix at face value. It has a habit of not looking at the whole picture. You have to prompt it to 'think bigger'.
@AI_Reveals I feel like most of the people that only care about using AI as a chatbot aren't using Claude.
They don't care as much about intelligence or coding. They care more about image and video generation.