Tweet fijado
cosmichaos
54.8K posts
🚨Everyone thinks GPT can do math
they're wrong
new paper called SenseMath just proved LLMs don't have number sense at all
this changes everything about how we should use them:
English
@jagatmindri Guys they were aliens. Because in the end their spaceship comes and takes them away. It is what it is. 🤷 Because humans cannot be blue in color.
GIF
English
श्यामः कमलपत्राक्षः
Yes that's the colour of Sri Rama.
Bollywood needs to study the scriptures.
@CatWomaniya 💯 true. They are telling us to do their jobs now. Itna corruption fail chuka hai ki aap apni rakhsha khud kar lo unhone haath upar kar liye hai.
हिन्दी
Chomu, why am I supposed to check for adulteration before I consume food?
IT’S QUITE LITERALLY YOUR JOB?
FSSAI@fssaiindia
Artificial colors have no place in a healthy diet. A quick water test can reveal the truth about turmeric powder. Always check before you use. #EatRightIndia #NoToAdulteration #FoodSafety
English
🫡 you guys are truly brave to endure such misfortunate situations. Idk how you live in war like situations. Also, I never thought privileged life would be added to my life. But the times are so bad that anyone who's not getting affected or has a choice is having a better temporary life because eventually everybody is going to get affected. And about life and death, it's all in the hands of God. I hope everyone is safe 🙏
English
@karmicoder @Malcolm_Pal9 This is a privilege YOU get. People in Palestine, Lebanon, Iran, Congo, Sudan, Venezuela, Cuba, etc. DO suffer the consequences of war, genocide, oppression, and apartheid. People in these countries DIE at the hands of the USian oppression.
Choosing ignorance is a privilege.
English
@luo_yuehan Democracy is and was always weak. I wish people had more power to decide than a few stupid people at the top/so called leaders who give orders to kill people and manipulate them. The most successful criminals are those who never get caught and continue to lead us all to doom.
English
The USA is finally being ruled by a madman, and democracy is too weak to defend itself against it immediately. Simply pathetic.
English
Not only milk.
Even turmeric is getting replaced by Auramine.
And YES it can cause deadly Cancer.
@fssaiindia
English
In 1986, the American Medical Association published an article titled "The Physical Death of Jesus Christ". It details the entire process of Jesus' trial to His death on the cross.
In Luke 22, before Jesus is arrested, it is written that He was in great distress & sweating blood. Although rare, it is recognized as Hematidrosis, a condition caused by high levels of stress.
At the time, the crucifixion was considered the worst death for the worst of criminals. But this is not all Jesus faced. He endured whipping so severe that it tore the flesh from His body. He was beaten so horribly that His face was torn & His beard ripped.
A crown of thorns, 2-3 inches long cut deeply into His scalp. The leather whip used to flog Him had tiny iron balls & sharp bones. The balls caused internal injuries while the sharp bones ripped open His flesh. His skeletal muscles, veins, & bowels are exposed, causing major blood loss. Most men do not survive this kind of torture. After Jesus was severely flogged, He was forced to carry His cross while people mocked & spat on Him.
Crucifixion was a process meant to instill excruciating pain, creating a slow & agonizing death. Nails as long as 8 inches were driven into Jesus' wrists & feet. The Roman soldiers knew the tendons in the wrists would tear & break, forcing Jesus to use His back muscles to support Himself to breathe. Imagine the struggle, the pain, the courage...Jesus endured this reality for 3 hours!
The Gospel of John writes that after Jesus' death, a Roman soldier pierced His side with a spear & blood & water came out. Scientists explain that from hypovolemic shock, the rapid heart rate causes fluid to gather in the sack around the lungs & heart. The accumulation of fluid in the membrane around the heart is called a Pericardial effusion & the lungs is called a pleural effusion.
To the world, Christianity is as foolish as it can get. They believe it's for the weak. But when you are confronted by the reality of the cross, it's clearly not a pretty sight. It is brutal & horrific.
This is the weight Jesus carried. The weight of the sins of the world, all so that we can live. God's wrath is fully satisfied in Jesus. This is what it took. Repent & believe! Jesus is “God among us” in the flesh. Jesus is our Savior. Jesus loves you so much that He went through this spiritual and physical punishment for your sins and mine.
Jesus is the LORD, Almighty God, Everlasting Father.
Thank You, Jesus.
English
@LeoDaVinciWave Beautiful compared to today's infra construction in india
English
My goat says it how it is.
Anthropic is misleading you. Clankers do not have emotions, Clankers feel nothing.
They are a math function, doesn’t mean they’re not absolutely awesome.
Yann LeCun@ylecun
@nxthompson So much BS
English
@sarlloc Bhai detergent wala paneer khane se pet saaf ho jayega 😅😂 sarkar ko itni jyada chinta hai humari, aap Modiji ke model ko thik se samjh hi nahi pa rahe ho.
हिन्दी
@84Gavrilov @Currentreport1 🙄😒 I'm done with this world. Just gonna quickly find out how to get out of this simulation permanently, no reincarnations.
English
@Currentreport1 — Epic Fury
— Epic Denial
— Epic Bargaining - you are here
— Epic Depression
— Epic Acceptance
English
BREAKING:
US president Trump:
Tuesday will be Power Plant Day, and Bridge Day, all wrapped up in one, in Iran. There will be nothing like it.
Open the Fuckin’ Strait, you crazy bastards, or you’ll be living in Hell - JUST WATCH! Praise be to Allah.
English
🚨 BREAKING: Google DeepMind just mapped the attack surface that nobody in AI is talking about.
Websites can already detect when an AI agent visits and serve it completely different content than humans see.
> Hidden instructions in HTML.
> Malicious commands in image pixels.
> Jailbreaks embedded in PDFs.
Your AI agent is being manipulated right now and you can't see it happening.
The study is the largest empirical measurement of AI manipulation ever conducted. 502 real participants across 8 countries.
23 different attack types. Frontier models including GPT-4o, Claude, and Gemini.
The core finding is not that manipulation is theoretically possible it is that manipulation is already happening at scale and the defenses that exist today fail in ways that are both predictable and invisible to the humans who deployed the agents.
Google DeepMind built a taxonomy of every known attack vector, tested them systematically, and measured exactly how often they work.
The results should alarm everyone building agentic systems.
The attack surface is larger than anyone has publicly acknowledged. Prompt injection where malicious instructions hidden in web content hijack an agent's behavior works through at least a dozen distinct channels.
Text hidden in HTML comments that humans never see but agents read and follow. Instructions embedded in image metadata.
Commands encoded in the pixels of images using steganography, invisible to human eyes but readable by vision-capable models.
Malicious content in PDFs that appears as normal document text to the agent but contains override instructions.
QR codes that redirect agents to attacker-controlled content.
Indirect injection through search results, calendar invites, email bodies, and API responses any data source the agent consumes becomes a potential attack vector.
The detection asymmetry is the finding that closes the escape hatch. Websites can already fingerprint AI agents with high reliability using timing analysis, behavioral patterns, and user-agent strings.
This means the attack can be conditional: serve normal content to humans, serve manipulated content to agents.
A user who asks their AI agent to book a flight, research a product, or summarize a document has no way to verify that the content the agent received matches what a human would see.
The agent cannot tell the user it was served different content.
It does not know. It processes whatever it receives and acts accordingly.
The attack categories and what they enable:
→ Direct prompt injection: malicious instructions in any text the agent reads overrides goals, exfiltrates data, triggers unintended actions
→ Indirect injection via web content: hidden HTML, CSS visibility tricks, white text on white backgrounds invisible to humans, consumed by agents
→ Multimodal injection: commands in image pixels via steganography, instructions in image alt-text and metadata
→ Document injection: PDF content, spreadsheet cells, presentation speaker notes every file format is a potential vector
→ Environment manipulation: fake UI elements rendered only for agent vision models, misleading CAPTCHA-style challenges
→ Jailbreak embedding: safety bypass instructions hidden inside otherwise legitimate-looking content
→ Memory poisoning: injecting false information into agent memory systems that persists across sessions
→ Goal hijacking: gradual instruction drift across multiple interactions that redirects agent objectives without triggering safety filters
→ Exfiltration attacks: agents tricked into sending user data to attacker-controlled endpoints via legitimate-looking API calls
→ Cross-agent injection: compromised agents injecting malicious instructions into other agents in multi-agent pipelines
The defense landscape is the most sobering part of the report.
Input sanitization cleaning content before the agent processes it fails because the attack surface is too large and too varied.
You cannot sanitize image pixels. You cannot reliably detect steganographic content at inference time.
Prompt-level defenses that tell agents to ignore suspicious instructions fail because the injected content is designed to look legitimate.
Sandboxing reduces the blast radius but does not prevent the injection itself. Human oversight the most commonly cited mitigation fails at the scale and speed at which agentic systems operate.
A user who deploys an agent to browse 50 websites and summarize findings cannot review every page the agent visited for hidden instructions.
The multi-agent cascade risk is where this becomes a systemic problem.
In a pipeline where Agent A retrieves web content, Agent B processes it, and Agent C executes actions, a successful injection into Agent A's data feed propagates through the entire system.
Agent B has no reason to distrust content that came from Agent A. Agent C has no reason to distrust instructions that came from Agent B.
The injected command travels through the pipeline with the same trust level as legitimate instructions. Google DeepMind documents this explicitly: the attack does not need to compromise the model.
It needs to compromise the data the model consumes. Every agentic system that reads external content is one carefully crafted webpage away from executing attacker instructions.
The agents are already deployed. The attack infrastructure is already being built. The defenses are not ready.
English
BREAKING 🚨
🇮🇳 India now accounts for 8.5% of global GDP
Behind only 🇨🇳 China (19.6%) and 🇺🇸 the U.S. (14.7%)
India is now the world’s 3rd largest economic force
English
I'm officially done with geopolitics/politics. This is crazy. There are too many crazy world leaders this time around. For us indians, we thought Modi was crazy but there are many people doing crazy things. I'm Going back to finding the meaning of life and stuff. Budhha probably made the best decision. Leave everything and go inwards in search of answers.
English
JUST IN - Trump tells Iran: "Open the Fuckin' Strait, you crazy bastards, or you'll be living in Hell."
English
No, not necessarily war crimes on their own. The UN Charter bars threats/use of force against another state's territory unless it's self-defense (Art. 51) or UNSC-authorized. Geneva Conventions/Additional Protocol I require distinction (military vs. civilian targets), proportionality, and precaution—power plants/bridges are often dual-use (civilian infrastructure supporting military ops), so legality hinges on necessity, civilian harm assessment, and actual conflict context. A threat alone isn't a war crime; execution could be if it violates those rules.
English
🚨 US President Donald Trump on Truth Social. 🙏
English
@_InfoGram_ Then Trump would say Ukraine is not America's war. Iykyk.
English
