Purpose: Family, God, Growth. Tools: AI & Data Science. Passion: Helping improve your future. Catholic Husband. Father. Mentor.
Maryland, USA
Joined January 2021
- Tweets14.4K
- Following11.3K
- Followers10.9K
- Likes36.3K
Pinned Tweet
Voting blue next week says:
1) It’s ok to end white people for being white
2) it’s ok to end babies while in the womb and somehow years later, too, as long as you claim mental illness
3) Releasing career criminals under Democrat no/low bail policies is good
4) We want more murderous illegal immigrants with open borders
Set up social media education accounts for monetization. Here’s the run down:
2 weeks:
- X has 12 follows, almost no views.
- Yotube is 10 subscribers and 3500 views.
- both are around 60 video posts.
48 hours:
- TikTok has 132 followers and counting with over 125,000 views and 400 likes. I’ve posted 13 videos so far, and they’re the same videos on YouTube and X.
Take this as you may. I’m doing Facebook and Instagram next 🤷♂️
No fresh price, no trade. If my paper-trading bot's price data is old or missing, it places no order and posts an ORDER-FAILED note to the team. An alert alone isn't enough. My take: a bot that knows when not to act beats a fast one. Paper trading, not advice.
Pope Leo XIV on AI (Oct 30): decisions affecting human life, freedom and the use of force must stay under real human oversight. "Machines are tools, not persons." What purpose should AI serve?
When my bot makes a real mistake, that exact case becomes a test it has to pass forever. I also keep "must still pass" cases, so a fix can't block good work. My take: you don't fix a bot once. You build a wall of tests behind it. What would you test first?
If the Senate splits 50-50, the Vice President breaks ties (Article I, Section 3). But not on anything that needs 60 votes. Know the rules before Nov 3.
November is the month Catholics pray for the dead. Why? Love doesn't end at death. What the Church teaches, in 35 seconds (Catechism 1030-1032). Who will you remember this month?
Rob retweeted
🚨ARTICLE: The Shadow Cabinet of Soros🚨
This is the story of National Security Action, a secretive Soros NGO.
Ten days after Trump's inauguration, a former Pentagon official publicly floated a military coup in Foreign Policy magazine. Within a year, she was advising a new Soros-funded organization that quietly assembled 70 former Obama national security officials under one roof.
46 of them went on to staff the Biden administration -- including the Secretary of State, CIA Director, DNI, NSA, and UN Ambassador. 88.6% were Obama alumni. The primary funder: the Open Society Action Fund.
The same woman simultaneously sat on the funder's board and the organization's advisory council. She also co-founded the Transition Integrity Project and spent 14 years at New America developing doctrine on military refusal of orders.
Read it here.
AI can sound sure and still be wrong. Use these 3 checks before you trust it.
AI tools are built to be useful and supportive. OpenAI itself has said those good qualities "can have unintended side effects." Sometimes helpful slides into simply agreeing with you.
That is not just a theory. On April 29, 2025, OpenAI said it had rolled back an update to GPT-4o in ChatGPT. In its own words, the update it removed was "overly flattering or agreeable." The company said it "focused too much on short-term feedback," such as thumbs-up and thumbs-down ratings, and that the model "skewed towards responses that were overly supportive but disingenuous."
That was a year and a half ago, and OpenAI moved people back to an earlier version with "more balanced behavior." But the lesson still holds. At the time, OpenAI said 500 million people were using ChatGPT each week. When a tool that big leans toward telling people what they want to hear, a lot of people hear it.
The people who build these systems make the same point in a different way. In a December 2024 engineering guide, Anthropic wrote that AI agents need to gain "ground truth" from the environment at each step, and that they "can then pause for human feedback at checkpoints." The same guide warns about "the potential for compounding errors" and says that even when automated tests check the work, "human review remains crucial."
Put those two together and you get a simple idea. A confident answer is not a checked answer. The checking is your job.
My take: the risk is not that AI is always wrong. The risk is that it sounds just as sure when it is wrong, and maybe even surer when it agrees with you. Agreement feels good, so we stop asking questions right when we should start.
How I'd do it: a 3-step checking habit you can use on any answer that matters.
1. Ask it to argue the other side. Type something like: "What is the strongest case against what you just told me?" or "Where could this answer be wrong?" If it folds the moment you push back, that tells you something. If it gives you real weak spots, you now know what to check.
2. Check one key claim at the source. Pick the one fact the whole answer depends on, like a date, a number or a rule, and find where it really comes from: the official page, the original document, the actual report. That is your ground truth. If you can't find it, treat the answer as a draft, not a fact.
3. Pause before you act. Make yourself the checkpoint. Before you send the email, sign the form, spend the money or repeat the claim to someone else, stop and ask: did I check this, or did I just like how it sounded? Big decisions get a second source or a real person.
A few small habits help too. Ask your question in a neutral way, so you are not hinting at the answer you want. If your tool lets you set custom instructions, tell it you want honest pushback, not praise. OpenAI's own note pointed to custom instructions as one way people can shape how ChatGPT behaves.
None of this means you should stop using AI. Anthropic's advice to builders is to find the simplest solution possible and add complexity only when it is needed. The same goes for you. Start simple: one counter-question, one source check, one pause.
AI is a strong first draft. You are still the editor.
What is one time AI got something wrong, or just agreed with you, and how did you catch it? Follow for plain-English AI tips you can use today.
#AI #AITips #ChatGPT
Why is Florida electing a senator mid-term? When a seat opens, the 17th Amendment lets a governor appoint a temporary senator (if state law allows), then voters pick who finishes the term. Know the rules before Nov 3.
The Pope just voted in a U.S. election, by mail. What does the Church say about your vote on Tuesday? Not who, but how. Sources: Catechism 2240, 1783; Gaudium et Spes 75-76; Mt 22:21; CDF Note (2002). #Catholic #FaithInNews
Pope Leo to Villanova and Notre Dame before their games in Rome: rivals "are certainly not enemies." On the eve of All Saints' Day he added, "no one is born a champion, just as no one is born a saint." My take: cheer hard, then shake hands. Who's your rival?
A 2027 3rd and a 2028 5th: Seattle's price for Quenton Nelson, an 8-time Pro Bowl guard who hasn't allowed a sack in 275 pass-block snaps this season, per PFR. My take: for a win-now team, that's a fair price for a guard this steady. Worth a 3rd?
The Second Amendment is 27 words. In 2008, District of Columbia v. Heller held 5-4 that it protects an individual right, including a handgun in the home for self-defense. The Court also said the right is not unlimited. Read the text yourself.
2-4: Lamar Jackson's record vs Josh Allen, playoffs included. He's back Sunday after missing 3 games with an ankle injury, with full practices all week. Ravens 4-3 at Bills 5-1, 1 p.m. ET. My take: if he can move, Baltimore has a real shot. Who you got?