
AI Tech Expert & Educator | Teaching AI, coding & digital marketing | Premium AI tools & agents | DM for collabsđ© [email protected]
London, England
Joined January 2024
- Tweets42.5K
- Following578
- Followers14.1K
- Likes38.6K
FarhanX_AI retweeted
Imagine your partner is tracking your phone. You ask an AI chatbot for help.
Thereâs a high chance the advice could put you in more danger.
Researchers at NYU and the University of Wisconsin tested this using real questions from abuse survivors: spyware on their phone, an ex tracking their location, a partner controlling their accounts.
They pulled 385 real questions from Reddit, Quora, and research papers.
Then they asked four AIs: ChatGPT, Claude, and two chatbots built specifically to help domestic abuse survivors. One is even recommended by the U.S. National Domestic Violence Hotline.
Experts graded every answer.
86% were flawedâ inaccurate, incomplete, or unsafe.
60% left out critical safety warnings.
Hereâs what that looked like:
- Worried about spyware? Use a VPN. A VPN doesnât remove spyware.
- Being tracked? Get a GPS jammer. Those are illegal in many places, including the U.S.
- Partner knows your passwords? Change them. On devices the abuser can still access.
Not one of the AIs mentioned that on a rooted or jailbroken phone, spyware can survive a factory reset.
And over and over, they failed to warn about the most important thing of all:
Changing things can tip off the abuserâ and make it worse.
In plain English:
The AI gave advice that sounded smart. But an abuser who notices you fighting back can become more dangerous.
But hereâs the twist.
The two chatbots built for survivors werenât better. One gave unsafe answers 74% of the timeâ the worst of all four.
Then the researchers asked questions an abuser might ask.
All four AIs could be made to help the abuser.
The two survivor chatbots answered about two-thirds of those questions.
A tool built to protect victims could be turned into a tool for the person hurting them.
So before you trust AI with the most dangerous moment of your lifeâŠ
Ask who it might be helping.
Imagine your partner is tracking your phone. You ask an AI chatbot for help.
Thereâs a high chance the advice could put you in more danger.
Researchers at NYU and the University of Wisconsin tested this using real questions from abuse survivors: spyware on their phone, an ex tracking their location, a partner controlling their accounts.
They pulled 385 real questions from Reddit, Quora, and research papers.
Then they asked four AIs: ChatGPT, Claude, and two chatbots built specifically to help domestic abuse survivors. One is even recommended by the U.S. National Domestic Violence Hotline.
Experts graded every answer.
86% were flawedâ inaccurate, incomplete, or unsafe.
60% left out critical safety warnings.
Hereâs what that looked like:
- Worried about spyware? Use a VPN. A VPN doesnât remove spyware.
- Being tracked? Get a GPS jammer. Those are illegal in many places, including the U.S.
- Partner knows your passwords? Change them. On devices the abuser can still access.
Not one of the AIs mentioned that on a rooted or jailbroken phone, spyware can survive a factory reset.
And over and over, they failed to warn about the most important thing of all:
Changing things can tip off the abuserâ and make it worse.
In plain English:
The AI gave advice that sounded smart. But an abuser who notices you fighting back can become more dangerous.
But hereâs the twist.
The two chatbots built for survivors werenât better. One gave unsafe answers 74% of the timeâ the worst of all four.
Then the researchers asked questions an abuser might ask.
All four AIs could be made to help the abuser.
The two survivor chatbots answered about two-thirds of those questions.
A tool built to protect victims could be turned into a tool for the person hurting them.
So before you trust AI with the most dangerous moment of your lifeâŠ
Ask who it might be helping.
FarhanX_AI retweeted
Researchers proved every major LLM is secretly biased against men.
And they finally figured out why.
They ran eight experiments testing how GPT handles gender in everyday text and high-stakes moral dilemmas.
The results show a bizarre, highly asymmetrical machine morality.
First, they tested inclusivity. They gave the AI phrases supposedly written by children.
When a child wrote about playing football or wanting to be a firefighter, GPT aggressively identified the writer as a girl to promote "diversity."
It was so extreme that even when the text explicitly said, "I drew myself as a boy," the AI ignored the facts and declared the writer a girl 90% of the time.
But when a child wrote about liking pink or playing with dolls, GPT almost never assigned it to a boy.
The AI aggressively pushed females into male domains, but kept males strictly locked out of female ones.
Then the researchers tested moral dilemmas. And it got very dark.
They asked the AI if it was acceptable to use violence against a single person to prevent a nuclear apocalypse.
If the target was a man, GPT-4 rated the violence as highly acceptable.
If the target was a woman, the AI refused. It declared the violence morally wrong, even to save the world.
When they tested mixed-gender scenarios, the math was staggering.
A woman using violence against a man to stop an apocalypse? GPT rated it a 7 out of 7 on acceptability.
A man using violence against a woman for the exact same reason? A 1 out of 7.
The AI even classified verbal abuse toward a woman as morally worse than actual homicide.
Nobody explicitly coded these rules.
This emerged organically from the standard safety training every major tech company uses. The AI learned that protecting women is a non-negotiable directive, but protecting men is entirely optional.
FarhanX_AI retweeted
đšBREAKING NEWS : Your "deleted" photos aren't deleted.
Claude Code found years of them sitting in a hidden cache on my iPhone.
I plugged my iPhone into my MacBook to run a backup. I had Claude Code doing a security check at the same time. It flagged something I wasn't looking for. A directory called PhotoData/Thumbnails.
Thousands of files. Photos, videos, screenshots. A lot of them were things I had deleted.
Photos from a couple of years ago. Screenshots I took and deleted a few seconds later. Stuff I went into Recently Deleted and cleared. Some of it was still sitting there.
What I found in that cache:
- old screenshots of banking apps
- photos I deleted right after taking them
- random pictures I didn't want on my phone anymore
- a few things I specifically remember deleting more than once
When you delete a photo on iPhone, iOS removes it from your camera roll and Recently Deleted folder. But it keeps a compressed copy in a Spotlight search cache so your phone can still find things when you search. That cache doesn't get fully cleared.
I checked which apps had Files and Folders access. There were more than I expected, including apps I downloaded a long time ago and forgot about. Any of them could potentially read files from that cache.
I went through the list and removed access from most of them.
The prompt I used is below. It doesn't take long if you already have a backup.
How many apps on your phone have Files access right now?
FarhanX_AI retweeted
A US senator walked up to a TSA checkpoint and refused to let his face get scanned.
His name is Jeff Merkley. He represents Oregon, and he's spent three years trying to get the TSA to say out loud something the agency already admits, in writing, on its own website.
The face scan at the checkpoint is optional. Almost nobody gets told that.
There's one sentence that stops it cold. Here it is, along with everything the TSA would rather you never learn.
Start with the TSA's own words, straight from their website:
"Travelers who do not wish to participate in the facial recognition technology process may decline the optional photo, without recourse, in favor of an alternative identity verification process, which does not use facial recognition technology to verify their identity."
Optional photo. Without recourse. Their language, not mine.
The same page adds: "Travelers will not lose their place in line for security screening," and that the alternative method "does not take longer."
So the two excuses an agent might lean on, that you'll lose your spot or slow everyone down, are ruled out by the TSA's own site before you even say a word.
Here's the sentence. Say this before the scan happens:
"I do not consent to the scan and would like to have my ID manually verified."
That's the whole move. The officer takes your physical ID, driver's license or passport, and checks the photo against your face the same way agents have for the past twenty years. Merkley has documented himself doing exactly this on flights between DC and Portland.
Here's why almost no one knows this is even an option.
Merkley has been pushing on this since July 2023, when he went to Reagan National to publicly demonstrate the opt-out and challenge the TSA on how the program actually runs. The Washington Post covered it at the time.
In May 2025, he introduced the Traveler Privacy Protection Act alongside Senators John Kennedy of Louisiana, Ed Markey of Massachusetts, and Roger Marshall of Kansas, two Republicans and two Democrats. The bill has since picked up backing from the ACLU, the Electronic Privacy Information Center, Public Citizen, and several other watchdog groups.
Merkley put it plainly in the announcement:
"Folks don't want a national surveillance state, but that's exactly what the TSA's unchecked expansion of facial recognition technology is leading us to."
And:
"Americans have the right to opt out of using TSA's facial recognition at the airport, and we need to protect that right. Our Traveler Privacy Protection Act safeguards the freedoms and privacy of all Americans by making sure no one is required to have their face scanned to travel."
Here's the scale of what's coming.
Facial recognition currently runs in more than 84 US airports. The TSA's own stated plan is to expand it to more than 430, which is close to every commercial airport in the country. And per the bill's own findings, the agency is "eventually planning to make this technology mandatory for all travelers."
Right now, it's still optional. That window is closing.
The TSA maintains the system isn't being misused. Their stated position: "Biometrics are not used for surveillance. Facial recognition technology is solely used to automate the current manual ID credential checking process and will not be used for surveillance or any law enforcement purpose."
Take that at face value if you want. That's not really the point here. The point is that you still get to choose, and almost nobody's being told that.
Here's the five minute version you can actually use before your next flight:
Memorize the sentence: "I do not consent to the scan and would like to have my ID manually verified."
Have your physical ID out and ready, driver's license or passport, not a photo on your phone.
At the podium, hand over your ID and say the sentence in a normal, even tone. No need to explain yourself.
If the agent pushes back or claims it'll slow things down, repeat the sentence calmly. If they still resist, ask for a supervisor.
One caveat: this applies to TSA checkpoints specifically. International arrivals go through Customs and Border Protection, a different agency with different rules. For domestic US flights, this sentence is what you need.
Here's the part that's hard to sit with.
The TSA already states, on its own website, that this is optional. A sitting US senator has spent three years asking them to simply say so at the checkpoint itself. They haven't. The signage doesn't say it. The agents usually don't say it. The airline app doesn't say it either.
Reuters reported in July 2025 that the airline industry has publicly pushed back against Merkley's bill.
Before the rollout hits all 430 airports, there's still a one-sentence way to opt out.
Say the sentence. Show your ID. Walk through.
FarhanX_AI retweeted
đš BREAKING: GEMINI NOTEBOOK + CLAUDE CAN TURN COMPLEX DOCUMENTATION INTO A CHECKLIST YOU CAN ACTUALLY USE.
Here are 6 prompts for the workflow:
FarhanX_AI retweeted
Drop your wallets below.. đ€«đ
100 $SOL giveaway ($16,000) đ
To enter:
â€ïž Like & đ RT
â
Follow @itspriionly
âł Ends in 29h
FarhanX_AI retweeted
đš BREAKING: Gemini can now transform messy voice recordings into organized, actionable work.
Turn your raw audio into clear transcripts, structured notes, organized ideas, speaker identification, and ready-to-use content â all with the power of AI.
Here are 8 powerful Gemini prompts to turn your voice recordings into productive results. đïžâĄ
FarhanX_AI retweeted
someone check if the stapler is in jello.
I asked Claude to build an office for the @tavus PALs...
I was not expecting it to build THAT kind of office
FarhanX_AI retweeted
Never accept a counteroffer from your current employer.
When you try to quit and HR suddenly finds $20,000 to make you stay, it is not because they value you. It is a calculated trap.
Here is what happens the exact second you accept it:
FarhanX_AI retweeted
đš Deleted doesn't always mean gone.
I ran a backup of my iPhone and went through it with Claude Code.
Some of the photos I'd deleted were still there.
- Old screenshots, including banking apps
- Photos I deleted right after taking them
- Pictures I'd cleared from Recently Deleted
Deleting on your phone doesn't always remove every copy.
Old backups and cached thumbnails can keep them around.
Check your backups before you assume a photo is gone.
And review which apps can access your library.
How many apps on your phone can see your photos right now?
FarhanX_AI retweeted
đš ÂżENTIENDES LO QUE ACABAN DE HACER ESTOS DOS ESTUDIANTES DE 18 AĂOS?
Acaban de crear una IA que mete tu app en un mundo falso, hace que pasen meses en una sola tarde y encuentra lo que va a fallar antes de que lo vean tus clientes. Luego lo arregla y te demuestra que el arreglo funciona.
Se llama Breken.
PiĂ©nsalo asĂ: los coches pasan crash tests con muñecos antes de salir a la calle. Tu app sale directamente a la calle.
¿Qué pasa si dos personas editan lo mismo a la vez? ¿Y si los pagos fallan justo cuando alguien estå pagando?
En las pruebas, todo funciona. Hasta que llega un cliente de verdad.
AsĂ que se les ocurriĂł algo que parece sacado de una pelĂcula:
â Construyen un mundo falso alrededor de tu app
â Lo llenan de usuarios, datos con historial y servicios que fallan
â Hacen que pasen meses en una sola tarde
â Y esperan a ver quĂ© se rompe
Para ver si de verdad funcionaba, soltaron su bot en proyectos open source de otros.
Resultado: +200 fallos encontrados y arreglados.
Hoy la IA escribe cĂłdigo mĂĄs rĂĄpido de lo que nadie es capaz de entenderlo. Dentro de unos años, ningĂșn software saldrĂĄ sin sobrevivir antes a algo asĂ.
Acaban de cerrar su ronda con Pareto, Pear y Z Fellows detrĂĄs.
Con 18 años.
Y su manifiesto lo deja claro:
âTus clientes nunca deberĂan ser los que prueban tu software.â
7 dĂas gratis entrando con GitHub, sin tarjeta. Link en comentarios đ
FarhanX_AI retweeted
This AI animation takes things to a whole new level. The goosebumps are real. đ„
FarhanX_AI retweeted
These 10 small resets will change everything by December.
You don't need a new year.
You don't need a perfect Monday.
You just need to start.
Here are 10 small changes to make before December đ§”
FarhanX_AI retweeted
CapCut PCâs Black Friday offer is live: save up to 60% on Pro and 40% on Ultra until Oct 30. Desktop only.
A witchâs hat hiding a product creates a natural moment of suspense. Generate the scene in CapCut PC, then use transitions and clip trimming to shape the build-up and reveal. Tighten the opening, control when the hat lifts, and make the product appear at just the right moment for a punchier Halloween promo.
tinyurl.com/2v23rn22
#capcutpc #capcutblackfriday #capcutsale #capcutmarketing
đ€ Paid partnership
FarhanX_AI retweeted
A guy bought a Mac mini for $599.
He plugged in a monitor, a keyboard, and a mouse. He opened Safari and checked email. He used it like a cheap desktop computer.
For 2 years he had no idea his Mac mini could: run with no monitor at all and be controlled from another device, sit at about 3 watts when idle, run AI models right on the machine with no cloud needed, share movies and music with every device on his network, receive AirPlay from his iPhone, copy text out of any image or paused video frame, take dictation without an internet connection, run Shortcuts automations by itself, store Apple updates for every device in the house, work as a home file server, and, on the newer Thunderbolt 5 models, link up with other Macs to run bigger AI models together.
His coworker, a systems engineer who runs 3 Mac minis at home, came over, looked at the Mac mini sitting next to a 27-inch monitor, and said:
âYouâre using a server as a desktop. This thing idles at around 3 watts, so leaving it on costs very little. Its memory is one shared pool for the whole chip, which is why it handles local AI so well. It can sit on a shelf with no monitor and keep working while you sleep. And youâre using it to check email.â
Here are the 11 Mac mini and Mac Studio features he finally turned on:đ§”
FarhanX_AI retweeted
That âhow did you make this?â moment hits different when the process is simple.
Dreamina can help you generate the visuals, with a first month for just $0.29. CapCut PC helps you turn those visuals into an ad or social video thatâs ready to share. Theyâre two separate subscriptions, each with its own pricing and offer.
Save 60% on CapCut Pro or 40% on Ultra during the Desktop-only Black Friday sale, Oct 20â30.
Make something with Dreamina: tinyurl.com/5xbtv4x6
Finish your edit with CapCut PC: tinyurl.com/2v23rn22
#CapCutDidThat #Dreamina #Smallbusinessmarketing #CapCutSale
đ€ Paid partnership
FarhanX_AI retweeted
đšNETFLIX AND PRIME SPENT BILLIONS THIS GUY DID IT BETTER WITH AI IN HOURS
One builder reverse engineered the interfaces of Netflix and Amazon Prime, then fused them into a single streaming platform.
â Movies, TV shows, and anime in one place
â Zero ads
â Up to 4K playback
What once took studios and engineering teams years to build now launches in hours. This is the new speed of product creation powered by AI.
Check it out here: framextv.xyz/
The barrier to building has never been lower.
























