GPTZero

Defending the internet from slop

Tag "@gptzeroai" in 🧵 to detect AI on X

[GPTZero @GPTZeroAI]

@alexcdot @mollycantillon Tag '@gptzeroai' to AI check any tweet

We are highly confident this text was AI-generated

View full analysis:
https://t.co/DQilOOyN23

[GPTZero @GPTZeroAI]

@edward_the6 @Hadas_Gold We are highly confident this text was human-written

https://t.co/LdPRcD09QE

[Hadas Gold @Hadas_Gold]

Here's what Nvidia's Jensen Huang sent to staff, admitting to affair:
"To my colleagues at NVIDIA,
I owe my family and all of you an apology.
I was unfaithful to my wife. My actions hurt her and our children. There is no excuse for what I did. I am deeply sorry.
Throughout my life, my family has given me love and support. They deserve my honesty and loyalty. Instead, I let them down and damaged the life we built together. I take full responsibility for what I did and the hurt I caused, and I accept the criticism that comes with it.
To my wife and family: I love you. I know an apology cannot undo what I have done. I hope our family will heal and that, over time, I can earn back your trust.
To my colleagues at NVIDIA, including those I have worked alongside for decades, and to our board, partners, and shareholders around the world: you deserve better from me. Leading NVIDIA is the privilege of my life. My responsibility is to serve you and be worthy of your trust. I have fallen short of that responsibility. I am fully committed to NVIDIA and will work to earn back your confidence.
Regarding the national security concerns raised in news reports, NVIDIA’s Board of Directors has initiated an independent investigation. I support the investigation and will cooperate fully. I love our country and our company. I am committed to protecting NVIDIA and supporting America’s national security. I am proud and grateful to be an American.
My family bears no responsibility for my choices. I do. Please give them privacy.
Jensen"

[GPTZero @GPTZeroAI]

@amckbell @RogerSeheult We are highly confident this text was human-written

https://t.co/mBorzqiJYt

[Roger Seheult, MD @RogerSeheult]

Check this out:
My patient today told me that when he asked his primary care, nurse practitioner about Ozempic to lose weight. She looked at him and said “you have two legs! use them.”

Well, he lives in Redlands and from that moment on, he started to walk around downtown Redlands at night.

He got to know all of the restaurant owners, and one of them even fills up his water flask each night

His wife made him a bright reflective jacket to wear when he walks. A lot of people think he’s a security guard walking around, but he goes every night.

He’s now up to 7 miles a night walking three hours.

He’s met a whole bunch of people that he’s become friends with.

People drinking think that he’s a cop and they stop drinking when they see him.

He’s become a friend of homeless people. One guy has seizures daily from when he was hit on the head with a crowbar. Because my patient has lost 30 pounds and six months from walking, he had some extra clothes that couldn’t fit him and he gave it to this homeless guy which thanked him profusely.

One day he saw that a pipe was leaking so he called it into the fire department and the next day when he walked by it was already fixed.

He walked by the library and he saw that someone had accidentally raised the American flag upside down. He went in and told them and they were very thankful and they fixed it immediately.

He’s now gotten to the point where if he doesn’t walk at night, he feels guilty. He’s sleeping better, he feels better, he’s lost weight, and he’s met a bunch of friends and changed people’s lives!

Wow.

I spent the whole clinic visit just hearing this guy‘s story because it’s amazing!

Imagine if all he did was go to the gym and walk on a treadmill.

(I got his permission to share this)

[GPTZero @GPTZeroAI]

@nathanclark @jasonlk @pangram We are highly confident this text was AI-generated

https://t.co/ys3DQkCWk1

[Jason ✨👾SaaStr.Ai✨ Lemkin @jasonlk]

Top 10 agent learnings from this week

1. Never tell an agent to "queue" work. Replit told us our approved fixes were "queued." It stopped overnight and built none of them. Now every prompt says "do it now."

2. We moved to cheaper models, and AI spend still went from ~$35 a day to $84. Call volume went up faster than the price per call came down.

3. I told the agent the old model HAD to go. A day later it was still running. I don't count it done until the call log shows zero calls in the last 24 hours.

4. Our AI fact-checker flagged 18 of 30 accurate notes because it treated opinions as factual claims. Every flag triggered a rewrite, so we paid twice for worse copy.

5. Each of the agent's daily reports said "no schema changes." The build broke anyway, on a table left over from an older change. The reports only checked each day's own work.

6. When an agent says data is missing, check the database first. This week the data was there all along, and the step that displays it was what was broken. Check the record before you pay to re-enrich anything.

7. Scout, our email agent, now acts on plain-English replies like "intro me" or "pause." It runs the same code as the buttons in the app, not a separate version.

8. One of our agents was answering emails sent to another agent's inbox for days before we caught it.

9. We canceled Canva after 12 years. We'd stopped using it a year earlier but kept paying, and two of us had separate team plans without knowing. We were paying about 4x for nothing.

10. 10K started as our AI VP Marketing. This week it pulled together SaaStr's financials, and it already runs invoicing, collections and commissions. We never planned any of that.

[GPTZero @GPTZeroAI]

@frea_bird @AaronBergman18 We are highly confident this text was human-written

https://t.co/yO78KY82Up

[Aaron Bergman 🔍 @AaronBergman18]

I love pangram but I feel like you could get an almost-as-good version working for *way* below their pricing

Not in a position to properly launch a new startup but still, dm/comment if anyone is interested

[GPTZero @GPTZeroAI]

@busgus @aidetectornews @dieMaehms @AmM55024055 We are highly confident this text was AI-generated

https://t.co/pCamb1Vvmj

[AI Detector News @aidetectornews]

@dieMaehms @busgus @AmM55024055 The sample size thing gets ignored constantly. Three sentences just don't carry enough signal for patterns in vocab and sentence variation to mean anything, but the score still gets quoted like it's a verdict. Short formal text is basically the worst case for these tools.

[GPTZero @GPTZeroAI]

@aiagentideas We are moderately confident this text was human-written

https://t.co/IHaKiG3IrJ

[Oystein | AI & Tech Entusiast @aiagentideas]

Your teacher might run your essay through an AI detector. Here's what nobody tells you: none of them are fully reliable.

I scored 7 AI detectors. What the scores say 🧵

[GPTZero @GPTZeroAI]

@edward_the6 @eddiegibbs @YankeeGunner We are highly confident this text was AI-generated

https://t.co/gQpfyGTRn4

[Eddie Gibbs @eddiegibbs]

What's the point of running an opinion piece through an AI detector? Presumably because challenging its arguments requires considerably more effort.

These tools are notoriously unreliable and cannot establish authorship. A percentage on a screen is hardly forensic evidence.

You've found a machine willing to agree with you. Congratulations. Now, have you anything to say about the football?

[GPTZero @GPTZeroAI]

@edward_the6 @QuinnyPig @clairevo We are highly confident this text was human-written

https://t.co/rl6Bfp0PtS

[Corey Quinn @QuinnyPig]

@clairevo Man, Pangram is terrible; it flagged this human-written tweet as AI generated for no good reason!

[GPTZero @GPTZeroAI]

@edward_the6 @kikollan We are highly confident this text was AI-generated

https://t.co/lqxIPCZbcH

[Kiko Llaneras @kikollan]

📚 Hemos encontrado una novela premiada con un 99% de TEXTO GENERADO con IA, según el detector Pangram.

'Donde el agua trabaja', del peruano David Silva Rojas, ganó un premio de 20.000 euros entre 2.204 manuscritos. Las bases prohibían IA.

Sería el primer caso de calibre en español 👇

[GPTZero @GPTZeroAI]

@edward_the6 @emollick We are highly confident this text was human-written

https://t.co/Ng0a5ElXQW

[Ethan Mollick @emollick]

The AI slopcomments are out-of-control on X but also remarkably easy to stop if the company wanted to?

I did a little experiment and found that combining as few as five comments and running them through Pangram is good enough to weed out the slop and not punish real humans.

[GPTZero @GPTZeroAI]

@edward_the6 @dharmesh We are highly confident this text was human-written

https://t.co/iDA7MbqqVJ

[dharmesh @dharmesh]

What I dislike more than AI Slop is what I call "AI Foam".

It's when generally smart, insightful people send you something that is clearly AI generated and lacks their usual sharp insight or point of view.

AI slop is easy to ignore because it's often just random people posting random stuff online -- often as content or comments.

AI foam is more insidious because these are people you *know* and respect and you know they're capable of critical thinking and clear communication.

I call it AI Foam because it lacks shape and substance.

[GPTZero @GPTZeroAI]

@edward_the6 @orlandosentinel We are highly confident this text was human-written

https://t.co/HEj1M3oVLY

[Orlando Sentinel @orlandosentinel]

The world’s leading AI writing detection software has flagged the majority of Henry Mack’s posts and much of his other written work as possibly written by AI. https://t.co/1yQv1At7Gq

[GPTZero @GPTZeroAI]

@edward_the6 @JESnowden We are highly confident this text was human-written

https://t.co/n8s7TtdE7u

[Jonathan Snowden @JESnowden]

An AI detection program claims the latest chapter in one of my books is 89% AI.

I'm starting to think these programs are not reliable. 🤣

You almost have to intentionally construct bad arguments and write weird sentences to pass these screenings.

[GPTZero @GPTZeroAI]

@edward_the6 @amplifiedamp We are highly confident this text was human-written

https://t.co/nOKewdml9O

[&. @amplifiedamp]

Call answered~

On my internal eval, Pangram failed to correctly detect 59% (±10%) of outputs generated by Opus 5.5 xhigh when augmented by my special MCP server using prompts from WildChat

If given multiple attempts, Gemini 4 Argon High achieves '100% human' on 58% of prompts.

[GPTZero @GPTZeroAI]

@_hmm3000_na1_B @yuki_shizu2140 We've already checked this tweet for AI!

[GPTZero @GPTZeroAI]

@_hmm3000_na1_B @yuki_shizu2140 We are highly confident this text was human-written

https://t.co/I50iRDOyGx

[GPTZero @GPTZeroAI]

@_hmm3000_na1_B @yuki_shizu2140 We are highly confident this text was human-written

https://t.co/I50iRDOyGx

[ゆき @yuki_shizu2140]

@hinaalona まさひこさん、YouTubeチャンネルを作るんですね。
「作ります」と短く言い切る一文に、これから始まる感じがあって、読むこちらも少し楽しみになりました。

[GPTZero @GPTZeroAI]

@fredrivett @nathanclark @pangram This text is too short for accurate AI detection. Please try with a longer tweet or thread!

[Nathan Clark @nathanclark]

How accurate is @pangram?

[GPTZero @GPTZeroAI]

@edward_the6 @sam_atis We've already checked this tweet for AI!

[GPTZero @GPTZeroAI]

@edward_the6 @sam_atis We are highly confident this text was human-written

https://t.co/5F5hgDU23e

[GPTZero @GPTZeroAI]

@edward_the6 @sam_atis We are highly confident this text was human-written

https://t.co/5F5hgDU23e

[Sam Glover @sam_atis]

This appears to be the first case of an actual Pangram 4 error I've ever seen in the wild.

[GPTZero @GPTZeroAI]

@edward_the6 @yoshispeninsula We've already checked this tweet for AI!

[GPTZero @GPTZeroAI]

@edward_the6 @yoshispeninsula We are highly confident this text was human-written

https://t.co/Q83oIZtMH4