Brady Bellini

I literally just got off a call with a customer and we were joking about how “silly” it was that we were all so hung up on model choice in the chat window… when in reality very few were actually selecting model based on data, but their own personal biases.

Are any apps launching with CLI's that an agent can control or is MCP still the way

You should value your privacy, especially when using AI. Excited to give this a try and will share my thoughts after testing more.

yOu'Re AbSoLuTeLy RiGhT dude

linear regress your p-values into submission

I'm honestly so excited for the jailbreak/hacker culture that will come from humanoid robots when they are mass market for consumers

I'm honestly so excited for the jailbreak/hacker culture that will come from humanoid robots when they are mass market for consumers

This more common than you think in cities. In Sacramento you can hire off duty police officers by paying their overtime. You can just go direct.

The hidden fragility of usage-based AI products: your entire revenue stream depends on someone else's infrastructure staying online.

Memory feature is more annoying than useful. It always wants to relate things back to memories. I never feel like I get a “clean” answer with memories enabled. Does feel dangerous bc of bias and convergence.

My AI stack shifted completely in the past 2 months: Before: ▸ChatGPT & 4.1 as default ▸Cline + Cursor for coding ▸Perplexity for quick research Now: ▸Claude & Sonnet 4 as default ▸Amp (with some Claude Code sprinkled in) for coding ▸ChatGPT reserved for o3 and deep…

So do we think OpenAI is still going to release a separate browser product, or is ChatGPT Agent the browser product?

I have such odd traffic patterns for my personal website. Is Ireland like a proxy hotspot for HTTP requests or something? Not sure why they would be so far above the rest.

Boomers should be banned from writing reviews. This is borderline spam lol

okay the naming is cool but also kinda scary