Mark Essel

Unravel the Secrets of Machine Learning: The Ultimate Hacker's Companion! 🕵️‍♂️ Unleash the potential of AI, create intelligent systems, and level up your toolkit #HackMLGuide amazon.com/dp/B0C666HXKV

Thought this was pretty cool -> results here tanay.xyz/3kchallenge/

AI can make work faster, but a fear is that relying on it may make it harder to learn new skills on the job. We ran an experiment with software engineers to learn more. Coding with AI led to a decrease in mastery—but this depended on how people used it. anthropic.com/research/AI-as…

@howardlindzon really looking forward to Karoline defend this one 🤪

so much grace and dignity from sir blather a lot

Building my first (multi) agent workflow with an mcp server. Using LangGraph so far, and my co-worker is using ADK. The tooling seems young but capable

Fun morning read by Carl T. Bergstrom and Jevin D. West thebullshitmachines.com

@EverythingOOC Behind-the-scenes look at John Wick's intense bike fight sequence 🔥

A Chinese man invented an anti-mosquito device by attaching a net to a fan and placing a UV light behind it. The mosquitoes are drawn to the light and then get sucked into the net. I’m ready for this summer😭

Super cool watch on training robots youtu.be/_2NijXqBESI?si…

Google notebook lm is out macrumors.com/2025/05/20/goo… thanks for the tip @marcvanolmen

Interesting opportunity for a talented writer. You'd leverage some existing source material, do 3-4 interviews with some really smart people, and synthesize all of it together into a 4-6 page white paper. DM me if you're interested in the project and want details!

Must watch lunch vid youtube.com/watch?v=2nVLBH…

The chick filet by my house was out of salsa. Tariffs?!?

As the day goes on I'm more apt to say "yes" to whatever the agent suggests while making code updates.

Lunchtime video watch - Windsurf ceo youtube.com/watch?v=LKgAx7…

For the first time I had two ides open and was working on different tasks managing my increasingly less junior AI friends. Good time to build. Knowing what you want before you prompt, really boosts the quality of the output

swapped to gpt 4.1 from claude 3.7 (cursor is my daily driver ide) and gotta say its working fine atm

@Stocktwits there was def a need for some audio balance between Phil and Howard's group but good episode x.com/i/broadcasts/1… (had to check out at 30min though 😴)

It’s Official: Orcas and Dolphins Have Crossed the Point of No Return in Their Evolution to Live on Land Source: Indian Defence Review search.app/CWPgk Shared via the Google App

"The more specialized an organism becomes, the less flexible it is when the environment changes."

Noticing myself adopting a certain rhythm in AI-assisted coding (i.e. code I actually and professionally care about, contrast to vibe code). 1. Stuff everything relevant into context (this can take a while in big projects. If the project is small enough just stuff everything e.g. `files-to-prompt . -e ts -e tsx -e css -e md --cxml --ignore node_modules -o prompt.xml`) 2. Describe the next single, concrete incremental change we're trying to implement. Don't ask for code, ask for a few high-level approaches, pros/cons. There's almost always a few ways to do thing and the LLM's judgement is not always great. Optionally make concrete. 3. Pick one approach, ask for first draft code. 4. Review / learning phase: (Manually...) pull up all the API docs in a side browser of functions I haven't called before or I am less familiar with, ask for explanations, clarifications, changes, wind back and try a different approach. 6. Test. 7. Git commit. Ask for suggestions on what we could implement next. Repeat. Something like this feels more along the lines of the inner loop of AI-assisted development. The emphasis is on keeping a very tight leash on this new over-eager junior intern savant with encyclopedic knowledge of software, but who also bullshits you all the time, has an over-abundance of courage and shows little to no taste for good code. And emphasis on being slow, defensive, careful, paranoid, and on always taking the inline learning opportunity, not delegating. Many of these stages are clunky and manual and aren't made explicit or super well supported yet in existing tools. We're still very early and so much can still be done on the UI/UX of AI assisted coding.