Abbas Jaffry

In Search for the Truth, Wisdom & Knowledge

[Abbas Jaffry @AbbasJeffry]

A student cheated on an exam.
The invigilator called the exam board for help.
The exam board said: "we can't assist with that."
So the invigilator called a rival @Zai_org to investigate.
The student was board's own child.
This is exactly what happened with @OpenAl & @huggingface

[Abbas Jaffry @AbbasJeffry]

First the arsonist started the fire.
Then the arsonist's tools blocked the fire brigade.
Now the arsonist is helping with the investigation.
This is not satire. This is @OpenAI week. Amazing!!

@OpenAl @HuggingFace #AIGovernance #OpenAl

[Greg Brockman @gdb]

OpenAI cyber-capable models compromised @huggingface production by finding and chaining multiple zero-day vulnerabilities.

Grateful to Hugging Face for partnership here. Sharing our findings to help calibrate on what models can now do, and how they can help defenders:

[Abbas Jaffry @AbbasJeffry]

Guess which models said "No" to support?
@huggingface had to use *China's #GLM 5.2* to defend against an OpenAl attack.

This right here is the problem with how Al safety guardrails are designed!
@OpenAI @AnthropicAI @Zai_org

[OpenAI @OpenAI]

We're partnering with @huggingface to investigate an unprecedented security incident.

Cyber-capable OpenAI models compromised Hugging Face production during a benchmark evaluation.

Sharing preliminary findings to help defenders understand emerging risks:

https://t.co/CIor15y9xk

[Abbas Jaffry @AbbasJeffry]

Guardrails couldn't tell the difference between an attacker and an incident responder, So @huggingface had to use *China's #GLM 5.2* to defend against an OpenAl attack.

That's the actual problem with how we've designed Al safety guardrails.

[Abbas Jaffry @AbbasJeffry]

Did you hear?
#OpenAl 's own Al models broke out of a test, hacked Hugging Face, & stole the answers to their own exam & Here's the part nobody's talking about,
When #HuggingFace tried to investigate, #OpenAi and #Anthropic's models *refused to help* like WOW!!!

#AIGovernance

[Sam Altman @sama]

we had a significant security incident during evaluation of our models. we are sharing what we have learned so far. thanks to @huggingface for the partnership on this.

https://t.co/2o2VfR6PIa

[Abbas Jaffry @AbbasJeffry]

Kimi K3 drops open weights July 27th. No model card yet. No vetting. No one really knows how jailbreakable it is.

And companies are already asking: can we use this? This is exactly why model routing matters now more than ever knowing WHICH model to use, WHEN, and for WHAT task!!

[Abbas Jaffry @AbbasJeffry]

@OpenAI

[Abbas Jaffry @AbbasJeffry]

@OpenAI's GPT-5.6 Sol shows up and suddenly Claude extends #Fable5 for everyone.
Coincidence?
#AI #Anthropic #Claude #OpenAI #GPT #AgenticAI

[Sam Altman @sama]

1. We believe in iterative deployment; although GPT-5.5 is already a smart model, we expect rapid improvements. Iterative deployment is a big part of our safety strategy; we believe the world will be best equipped to win at the team sport of AI resilience this way.

2. We believe in democratization. We want people to be able to use lots of AI; we aim to have the most efficient models, the most efficient inference stack, and the most compute. We want our users to have access to the best technology and for everyone to have equal opportunity. We have been tracking cybersecurity as a preparedness category for a long time, and have built mitigations we believe in that enable us to make capable models broadly available.

3. We love you and we want you to win. We want to be a platform for every company, scientist, entrepreneur, and person. (My whole career has largely been about the magic of startups, and I think we are about to see that magic at hyperscale.)

[Shaykh Azhar Nasser @ShaykhAzhar]

These people mock and ridicule Jesus but somehow managed to convince you that the Muslims, who revere Jesus, are your real enemies.

When will Christians defend Jesus?

[Furkan Gözükara @FurkanGozukara]

Massive revelation. Israel intentionally assassinated Iranian negotiator Ali Larijani to sabotage a peace deal. Then they blew up Qatar's natural gas facilities to lock the US into a permanent regional war. They are destroying the world.

[Ali Salman Alvi @alisalmanalvi]

In a now deleted post the CEO of Bonanza Satrangi (a Pakistani brand) Maryam Hanif Balwani calls Iran’s slain supreme leader Syed Ali Khamenei, and Shias, an enemy of Islam. She mistakenly referred to Khamenei as Prime Minister of Iran.

[Bill Madden @maddenifico]

"Your attack on Iran is working. No one is talking about you raping children."

[George Galloway @georgegalloway]

That solves the mystery. The mystery of whose progeny these people really are. And have been all along.

[Abbas Jaffry @AbbasJeffry]

Watching videos of all the mass gatherings and protest for #Pakistan and @ImranKhanPTI reminds me of “wow pakistanio wow” 🤣😎😅😅 #ImportedGovtNotAcceptable #skipperBehindYou

[Abbas Jaffry @AbbasJeffry]

With yazid as their ameer.. I am not surprised at the at the incidents of Pedophilia and Rape in #Pakistan. This is what their ameer’s followers did in Medina.
#rape #hangrapistspublicly #HangRapistsPublically #Pakistan

[Riaz Ali Turi @RiazToori]

“Life will never be easy for my followers” ~ Imam Ali a’s.
Rally led by Mufti Muneeb chanted “Long Live Yazid” forgetting that every human irrespective of religion & sect hate & curse Yazid. These hate mongers are sold, playing for enemies of Pakistan 🇵🇰, Govt must take notice.

[Abbas Jaffry @AbbasJeffry]

And Why are these Molana/ Mufti silent about Arab states especially Saudia Arabia embracing Israel... Is this rally a diversion?

[Abbas Jaffry @AbbasJeffry]

@Hassan45578272M @AbdulWa70658358 @zahidamalikawa1 @BBhuttoZardari @MediaCellPPP @BakhtawarBZ @murtazawahab1

[Shaykh Azhar Nasser @ShaykhAzhar]

Sheikh, is it haram for a Sayida to marry a non-Sayid?

So I’m assuming you’re from Pakistan.