MrComputerScience

I feel like Bostonians made a deal with the devil for a beautiful weekend. The conversation went like this: Bostonians: We request a beautifully warm January in Massachusetts. Devil: You must offer a great sacrifice. Bostonians: Hmmm. A superbowl ring? Devil: Very well.

@NatashaCL7 I don't have a dislike button. Oh well. Nobody listens to me anyway, lolz.

Oops, I forgot we’re suppose to dislike comments we don’t like now instead of just hiding the reply and blocking them.

What happens when you delete a ChatGPT conversation? Most people assume it’s gone for good. But it’s not that simple. In this piece, I unpack how OpenAI’s retention policy works, what gets deleted, and what lingers. Details below. FREE. ↓ #ChatGPT #AI #Privacy

⟶ Think your IT team can't see you using ChatGPT/Claude/Grok at work? As someone who has worked as a sysadmin for years, the reality might shock you. I broke it down + shared real safeguards to keep your personal AI use private at work. Read below for free. ↓ #AI #Privacy

AI-generated war footage is flooding social media during a real geopolitical conflict. Fake missile strikes. Fake explosions. Fake casualties. Millions watched. Read free below. ↓ #AI #Deepfakes

@_atlantx @PithyCyborg Thank you so much! I published a bunch of new technical AI articles. But never sent them to my newsletter. So, I didn't think anyone would ever see them. I am very honored that you found them + enjoyed reading!!!! Cordially, Mike D

@MrComputerSci @PithyCyborg These articles just happened to become my favorite thing to read now on the internet and about AI. Really a fan of what you do!

Most RAG systems do not break. But they DO rot. You ship with perfect retrieval. Six weeks later the "vibes" are off. The accuracy craters. This is the Silent Killer of AI infrastructure. Embeddings Drift. Here is the FREE guide on stopping the decay. #AI #RAG Details below. ↓

@MrComputerSci @PithyCyborg Another masterpiece arrived!!

@_atlantx @PithyCyborg Lol. Hell yeah! THANK YOU SO MUCH. This was just a random article I published that I will not share anywhere really, lol. (Publishing these for SEO purposes.) So, it means a lot that at least someone I know likes it. :) Hope you have a beautiful day. Cordially, Mike D

“The people who are crazy enough to think they can change the world are the ones who do.” — often attributed to Steve Jobs

Just published the latest issue of Pithy Cyborg. Latest #AI news I'm reporting: 1. Iran war deepfakes rack up millions of views. 💣 2. ServiceNow CEO warns college grads face 30% unemployment. 👀 3. New study says AI promotes delusions in prone users. Read for free. ↓ #Tech

"The real hero is always a hero by mistake; he dreams of being an honest coward like everybody else."

⟶ I just published a brand new article that talks about what it means when an AI company claims they "don't train on your data." Are they lying? My article spells in out in detail including four SHADY reasons they use your data anyway lol. Details below. FREE. ↓ #AI #Privacy

I asked Grok to grade my code + breakdown on semantic rot. The verdict? "This is 1337 as hell." I'm giving it all away for free. Screenshot in the replies. #Grok #1337 #RAG #LLM #Science #Architecture #Coding #AmWriting

If your retrieval quality is dropping but your logs show zero errors... you have a "Silent Failure" on your hands. Most coders will monitor latency. But almost no one monitors semantic shift. Here is the 2026 playbook for catching drift before your users do. #SoftwareEngineering

Your RAG system scored 94% in testing. Then it hit production and fell apart. 😬 I just published a new tech guide breaking down why this happens. It's NOT a bug. The way most teams test RAG is fundamentally different from how it actually gets used. Details below. ↓ #AI #RAG

How I set up a 30-day AI adversarial thinking experiment (and what actually made it work): → One rule before each session: state your position clearly, then say "argue against this, go for the weak spots, don't soften it" → You must restate the counterargument in your own words before you're allowed to dismiss it → Keep a running note of which objections you dismissed and why, you'll see the pattern fast → Do it for decisions that feel already-made, those are the most interesting ones The AI won't care about your feelings about the decision, and that's the whole point of the experiment.

AI makes getting answers easy. That’s not the hard part. The hard part is trusting where they came from. At scale, output isn’t the question. Your data is.

The elite salad garden AI prompt is my personal favorite. :)