Hi, I am rather new.
I started last month to get interested in LLM and entered the “Stuff AI into everything, and use it fore anything ignore OWASPLLM09 and see what works” hyperfocus zone. Experimented with the user side of things, thought I accidentally got Claude to reveal critical security infrastructure to me (it didn’t but I must have looked like a monkey to the hacker1 Analyst that reviewed my report…), and have now been spitballing ideas to train my own Model. The Initial Challenge: see how far I can get learning about AI, the inner workings of an LLM, and everything related, just by asking the very thing I am learning about nicely. I have no clue how far I actually got, but even though I am not 100% at the end of the road (or one of the 30 Roadmaps I let Gemini curate for me) I am nearing the end of my Experiments theory wise and starting my Experiments on getting into the Code side of things, again the challenge consists of ignoring LLM09 and just letting the Ghosts in the machine take the wheel with me merely orchestrating it, changing numbers, but letting the Codemachines do the actual coding. I hope my understanding developed enough to actually pull that part through. I know it developed enough to know I am probably doing a stupid, but hey. It’s fun anyways and I know how to navigate those probability machines a bit now… also I know what confabulating is… again, awkward situation, not cool Claude…
I thought I’d just say hello as I’m probably gonna be frequenting the forum more often from now on. As I can already see that experimenting with the Code is gonna be REALLY frustrating if you’Re not allowed to write much yourself (not that I could… I can barely read it.)
So yeah, Hi I am probably the most enthusiastic dumdum on here but hope We’ll get along nicely.