Ramblin' & Amblin' (2026-10-22): Another AI tip here is I taught mine

Another AI tip here is I taught mine to mine the snake in a can trick is that there's a snake in a can is a old prank trick that would people play on each other it would be a novelty gag got a can of peanuts you want to open this can of peanuts and hand it to your buddy the buddy opens it and it's not peanuts it's the snake coiled up and you're gonna shoot out of your face and scare you I mean everybody's gonna be having a good time except you but yeah the snake in a can so what I did is I said okay when approaching a dependency or a new API or being presented with new information to give it this snake in a can test one is remember that there sometimes are surprises inside of a can and that there's a few things you can do to check and prepare for whatever might be inside of that can you've been handed a can and you can inspect it visually is it a real can or is it a make-believe can inspect the label inspect the packaging has been appropriately sealed is the packaging uh hole stuff like that and then uh um so you've inspected it all right now uh it's past the inspection you're going to uh give a little give a little shake give a little shakes feel feel what's in there uh does it feel like you'd see feel uh bean shifting doesn 't necessarily mean they're uh the beans you I mean the peanuts you want but uh do you feel them shifting around in there all right do you want to inspect further okay um are the peanuts um actually uh as i said actually peanuts or maybe it's uh macadamias because uh you know you were told you were gonna get peanuts but you got macadamias might look good might look like you what you want but it's not what you were told you were gonna get so that's also something to be aware of and then there's also um are they are the peanuts um good you know on the peanut uh but uh you can end up on a long song with with a single peanut um and it doesn't end well so working through these kinds of uh guidances and seeing the ai when it is approaching something novel like a dependency I've seen this like like to tell you exactly when I saw it it'd be hard but um but to see it is surreal that you put something in your prompt and then you see this reasoning self-reasoning conversation and the thinking uh log and uh um it's it's been helpful and at least knowing that um it acts as a guard when properly engaged the uh ai will go through a checklist of of of guards but I definitely think that the fact that they're not trying to be descriptive I'm not saying check for you know cves check for uh this specific type of vulnerability um that that does get accrue around that space but the um guidance to kind of just let the ai figure out what it can do and how to approach a problem is how I i see that is that's uh I've always treated the tool it's like intent I i'm pushing the intent into it uh not so much uh specifics um about strategy but you know teaching it like I teach a mentor somebody I'm pairing with and I'm talking about how I'm approaching something and I'm I'm reasoning with it and then seeing you know the tool for for whatever it is um intelligent conscious or not uh using these these you kind of metaphorical structures to uh to uh flexibly apply what at least on that surface feels like a reasonable uh um chain of of of actions that are based off of this kind of abstract um scenario and that it it knows to apply that against all of the uh context that it's operating against again this is uh going to be uh just my personal experience all right thanks bye