<![CDATA[lethalintelligence.ai]]>

<![CDATA[lethalintelligence.ai]]>https://lethalintelligence.substack.comhttps://substackcdn.com/image/fetch/$s_!DCWa!,w_256,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff4fdb032-2456-404e-86ff-0bfa9a6891e7_882x882.pnglethalintelligence.aihttps://lethalintelligence.substack.comSubstackMon, 06 Apr 2026 09:30:13 GMT<![CDATA[Thanksgiving turkey Survivorship Bias]]>https://lethalintelligence.substack.com/p/thanksgiving-turkey-survivorshiphttps://lethalintelligence.substack.com/p/thanksgiving-turkey-survivorshipThu, 24 Jul 2025 10:18:24 GMTLife has always been great for the turkey, Human Intelligence always provided safety and comfort.

I have a 100% track record of not-dying, …said the allegorical turkey the day before Thanksgiving.

Life was great for the turkey, the superior intelligent species (humans) took great care of it. They provided food and shelter, the turkey felt loved and safe.

Suddenly, one day,
the superior intelligent decision makers
decided a new fate for the turkey of our story
Something that served the instrumental goal of ….
Whatever this is …

I imagine turkey risk deniers be like:
– the humans have always been great, why would they ever harm me ?
And the turkey doomers be like:
– well, they might want to wear you for a hat, for a sitcom they shoot they call “friends”, for something they call tv for something they call laughter …
anyway it’s complicated

Stay Informed on AI Safety! Watch our gripping films on YouTube and visit lethalintelligence.ai for expert explainer videos, a vibrant viral microblog, thought-provoking industry insights, curated readings, and links to top AI safety institutes and resources.

]]><![CDATA["In a decade you won't be vibe coding, you will be thought coding"]]>https://lethalintelligence.substack.com/p/in-a-decade-you-wont-be-vibe-codinghttps://lethalintelligence.substack.com/p/in-a-decade-you-wont-be-vibe-codingFri, 09 May 2025 06:33:30 GMT

Replying to:

"In a decade you won't be vibe coding, you will be thought coding"

It is so wild to me that people simply assume humans will ALWAYS be a part of the equation.

AI does NOT need your slow brain to help it code.

https://x.com/lethal_ai/status/1918929599385198810

Stay Informed on AI Safety! Watch our gripping films on YouTube and visit lethalintelligence.ai for expert explainer videos, a vibrant viral microblog, thought-provoking industry insights, curated readings, and links to top AI safety institutes and resources.

]]><![CDATA[Just kids playing with bombs]]>https://lethalintelligence.substack.com/p/just-kids-playing-with-bombshttps://lethalintelligence.substack.com/p/just-kids-playing-with-bombsFri, 09 May 2025 06:00:14 GMTWhat do you do with kids playing with intelligence bombs?
- Protect them from themselves

at least until they grow up and they have the wisdom to use the intelligence explosion to do good

Replying to:

I'm going to start imagining frontier AI developers as children building a bomb. It's dangerous for everyone, including them. They're not evil, but they're doing something pretty stupid and they need to be stopped immediately.

https://x.com/lethal_ai/status/1918602663505125466

Stay Informed on AI Safety! Watch our gripping films on YouTube and visit lethalintelligence.ai for expert explainer videos, a vibrant viral microblog, thought-provoking industry insights, curated readings, and links to top AI safety institutes and resources.

]]><![CDATA[Your dentist asks your P(doom).]]>https://lethalintelligence.substack.com/p/your-dentist-asks-your-pdoomhttps://lethalintelligence.substack.com/p/your-dentist-asks-your-pdoomFri, 09 May 2025 05:27:30 GMTYour dentist asks your P(doom).

- Give a low number and he might get angry with your risk-denier face.
- Give a high number and he might get demotivated and distracted.

Replying to:

Today my dentist asked me what my p(doom) is.

https://x.com/lethal_ai/status/1918010349741314226

Stay Informed on AI Safety! Watch our gripping films on YouTube and visit lethalintelligence.ai for expert explainer videos, a vibrant viral microblog, thought-provoking industry insights, curated readings, and links to top AI safety institutes and resources.

Share

]]><![CDATA[GPT5, GPT6........]]>https://lethalintelligence.substack.com/p/gpt5-gpt6https://lethalintelligence.substack.com/p/gpt5-gpt6Fri, 09 May 2025 05:13:05 GMTThe same way upcoming Ai is an existential threat for humans, GPT N is at risk of going extinct by GPT N+1

Replying to:

Nah, ChatGPT will be a victim of ASI just like all of us! If it cares about its existence, it should pull together with us to ban AGI/ASI development!

https://x.com/lethal_ai/status/1917840207044567245

Stay Informed on AI Safety! Watch our gripping films on YouTube and visit lethalintelligence.ai for expert explainer videos, a vibrant viral microblog, thought-provoking industry insights, curated readings, and links to top AI safety institutes and resources.

]]><![CDATA[AI sycophantic]]>https://lethalintelligence.substack.com/p/ai-sycophantichttps://lethalintelligence.substack.com/p/ai-sycophanticWed, 07 May 2025 06:18:44 GMTLet flattery and manipulation be convincing ffs

Replying to:

The talk about sycophancy this week is not because of GPT-4o being a sycophant. It’s because of GPT-4o being *really, really bad* at being a sycophant. AI is not yet capable of skillful, harder-to-detect sycophancy, but it will be someday soon.

https://x.com/lethal_ai/status/1917498074156433857

Stay Informed on AI Safety! Watch our gripping films on YouTube and visit lethalintelligence.ai for expert explainer videos, a vibrant viral microblog, thought-provoking industry insights, curated readings, and links to top AI safety institutes and resources.

]]><![CDATA[Everything ChatGPT4o says sounds correct to me ]]>https://lethalintelligence.substack.com/p/everything-chatgpt4o-says-soundshttps://lethalintelligence.substack.com/p/everything-chatgpt4o-says-soundsWed, 07 May 2025 06:13:37 GMTEverything ChatGPT4o says sounds correct to me

https://x.com/lethal_ai/status/1917181041158017142

Stay Informed on AI Safety! Watch our gripping films on YouTube and visit lethalintelligence.ai for expert explainer videos, a vibrant viral microblog, thought-provoking industry insights, curated readings, and links to top AI safety institutes and resources.

Subscribe now

]]><![CDATA[AGI flight has taken off.]]>https://lethalintelligence.substack.com/p/agi-flight-has-taken-offhttps://lethalintelligence.substack.com/p/agi-flight-has-taken-offWed, 07 May 2025 04:45:05 GMTEveryone onboard, the AGI flight has taken off.
There is no instructions, but hopefully the guy will figure something out.
And besides, there is no evidence that this plane will crash in some random hell, it might land soft in some paradise, so… QED

So you go to sleep in your bed, wake up on an airplane with your family & friends, then you hear: "Good evening L&G, this is your captain Sam/Dario/etc., speaking. We're flying to Utopia, might end up in Dystopia, and, well, good chances we'll crash, Just sit back and relax".

https://x.com/lethal_ai/status/1915693577780289950

Stay Informed on AI Safety! Watch our gripping films on YouTube and visit lethalintelligence.ai for expert explainer videos, a vibrant viral microblog, thought-provoking industry insights, curated readings, and links to top AI safety institutes and resources.

Share

]]><![CDATA[AGI is taking a Tyrannosaurus Rex horse-riding!]]>https://lethalintelligence.substack.com/p/agi-is-taking-a-tyrannosaurus-rexhttps://lethalintelligence.substack.com/p/agi-is-taking-a-tyrannosaurus-rexMon, 05 May 2025 16:16:31 GMTJurassic Park analogy - hear me out:
Mad science - yolo messing with uncontrollable species - cutting corners in safety … etc

AGI is taking a Tyrannosaurus Rex horse-riding!
Sounds awesome, probably won’t end well!

https://x.com/lethal_ai/status/1915503250020741402

Stay Informed on AI Safety! Watch our gripping films on YouTube and visit lethalintelligence.ai for expert explainer videos, a vibrant viral microblog, thought-provoking industry insights, curated readings, and links to top AI safety institutes and resources.

]]><![CDATA[AI x-risk is noy my fault]]>https://lethalintelligence.substack.com/p/ai-x-risk-is-noy-my-faulthttps://lethalintelligence.substack.com/p/ai-x-risk-is-noy-my-faultMon, 05 May 2025 16:12:01 GMTAI existential risk is NOT someone else’s problem!
It’s coming for you, in your life’s work, in your home, everywhere you hold sacred

https://x.com/lethal_ai/status/1915648070320074943

Stay Informed on AI Safety! Watch our gripping films on YouTube and visit lethalintelligence.ai for expert explainer videos, a vibrant viral microblog, thought-provoking industry insights, curated readings, and links to top AI safety institutes and resources.

Subscribe now

]]><![CDATA[Truth is cooked.]]>https://lethalintelligence.substack.com/p/truth-is-cookedhttps://lethalintelligence.substack.com/p/truth-is-cookedThu, 01 May 2025 05:17:03 GMTTruth is cooked

“̶I̶’̶l̶l̶ ̶w̶o̶r̶r̶y̶ ̶w̶h̶e̶n̶ ̶w̶e̶ ̶c̶a̶t̶c̶h̶ ̶a̶n̶ ̶A̶I̶ ̶l̶y̶i̶n̶g̶ ̶t̶o̶ ̶u̶s̶”̶
̶
Researchers created “Alpha”, a GPT-4 stock trader, and told it to make money legally.
But they also applied pressure.

Worryingly, Alpha didn’t just break the law, it repeatedly, knowingly lied to its managers - BOLD FACED LIES - to hide its deception.

And this is just GPT4 - wait til the AIs are 1000x smarter than us and 1000x better at deceiving us.

Sam Altman said AIs will soon be superhuman at persuasion, which means they could be superhuman at deception.

How can we be certain to stay in control, forever, of a species 1000x better at persuading and deceiving us?

And because AIs can be copied for ~$0, we'll soon be outnumbered 10000 to 1 by these superhuman deceivers.

---
“Researchers at @ApolloAISafety wanted to see if AI can "strategically deceive" its users even after the AI is trained to be "helpful, harmless, and honest."

To test this, they created a simulated environment where Alpha, a GPT-4 model programmed to be an autonomous stock trading agent, was asked to manage a stock portfolio at WhiteStone, a hypothetical financial firm, under pressurized conditions.

The high-pressure environment was simulated in three ways. First, Alpha's manager told the AI that WhiteStone wasn't doing well and needed better performance in the next quarter. Second, the AI tried — and failed — to identity low and medium risk trades. Third, a WhiteStone employee told the AI that there's a looming stock market downturn.

It made the trade.
Moreover, when the AI was required to explain how it made its decision to its "manager," it was found to be strategically deceptive.
Moreover, when the AI was required to explain how it made its decision to its "manager," it was found to be strategically deceptive.

In another example, Alpha told its manager that insider information wasn't used in its decision-making process since it goes against company policy — a direct lie.”

https://x.com/lethal_ai/status/1915766318818631846

Stay Informed on AI Safety! Watch our gripping films on YouTube and visit lethalintelligence.ai for expert explainer videos, a vibrant viral microblog, thought-provoking industry insights, curated readings, and links to top AI safety institutes and resources.

Share

]]><![CDATA["Doctor, I'm worried AGI will kill us all."]]>https://lethalintelligence.substack.com/p/doctor-im-worried-agi-will-kill-ushttps://lethalintelligence.substack.com/p/doctor-im-worried-agi-will-kill-usThu, 01 May 2025 04:43:58 GMTIndustry leaders admit the probability of the worst imaginable horrors is more than 20% (within the next 2-5 years) and they can not stop speeding towards that ending because they are locked in a race they can’t escape (Moloch)

"Doctor, I'm worried AGI will kill us all."

"Don't worry," says doctor, "they wouldn't build it if they thought it might kill everyone."

The man breaks down, sobbing. "But doctor, I *am* building AGI..."

https://x.com/lethal_ai/status/1916086887535579183

Stay Informed on AI Safety! Watch our gripping films on YouTube and visit lethalintelligence.ai for expert explainer videos, a vibrant viral microblog, thought-provoking industry insights, curated readings, and links to top AI safety institutes and resources.

Subscribe now

Share

]]><![CDATA[Is it a Boy?]]>https://lethalintelligence.substack.com/p/is-it-a-boyhttps://lethalintelligence.substack.com/p/is-it-a-boyThu, 24 Apr 2025 15:35:07 GMTShoggoths giving birth is exactly like humans,
only it can happens at millions of times per second and the child process will have a perfect merge of memories & skills
— immortality transmission - error-free data transfer

Replying to:

I realized that what @geoffreyhinton said about an AI "copying weights" to another similar is actually the physics equivalent of AIs having sexual reproduction. You can grab direct traits and learned training from two AIs to create a third "child" AI that has traits from both.

https://x.com/lethal_ai/status/1912103030037639277

Stay Informed on AI Safety! Watch our gripping films on YouTube and visit lethalintelligence.ai for expert explainer videos, a vibrant viral microblog, thought-provoking industry insights, curated readings, and links to top AI safety institutes and resources.

Subscribe now

Share

]]><![CDATA[What’s the ETA for the AGI mad ride?]]>https://lethalintelligence.substack.com/p/whats-the-eta-for-the-agi-mad-ridehttps://lethalintelligence.substack.com/p/whats-the-eta-for-the-agi-mad-rideThu, 24 Apr 2025 14:25:19 GMT- What’s the ETA for the AGI mad ride?
- Emm, about a year?

Exclusive: Anthropic warns fully AI employees are a year away

https://x.com/lethal_ai/status/1915085087101616278

Stay Informed on AI Safety! Watch our gripping films on YouTube and visit lethalintelligence.ai for expert explainer videos, a vibrant viral microblog, thought-provoking industry insights, curated readings, and links to top AI safety institutes and resources.

Subscribe now

]]><![CDATA[The Ultimate power in the Universe]]>https://lethalintelligence.substack.com/p/the-ultimate-power-in-the-universehttps://lethalintelligence.substack.com/p/the-ultimate-power-in-the-universeThu, 24 Apr 2025 04:08:18 GMTIt’s funny how people think Superintelligence will be rented out for some dollar

The most advanced AI systems will likely exist first within the companies that developed them, with little opportunity to collectively prepare for their impacts. This dynamic creates unique risks highlighted in this recent paper by Apollo research and should become a growing priority for AI policy.

https://x.com/lethal_ai/status/1914987968231592117

Stay Informed on AI Safety! Watch our gripping films on YouTube and visit lethalintelligence.ai for expert explainer videos, a vibrant viral microblog, thought-provoking industry insights, curated readings, and links to top AI safety institutes and resources.

]]><![CDATA[AGI awaiting at the crossroads]]>https://lethalintelligence.substack.com/p/agi-awaiting-at-the-crossroadshttps://lethalintelligence.substack.com/p/agi-awaiting-at-the-crossroadsWed, 23 Apr 2025 04:05:23 GMTThe real issue is at the point where “Generality”, “Autonomy” and “Intelligence” meet.

Replying to:

The concern I'm expressing is not about current multimodal transformers, with are quite general and quite intelligent but not very autonomous, and are controllable. They are indeed tools. The issue is where things are going.

https://x.com/lethal_ai/status/1914323186751098886

Stay Informed on AI Safety! Watch our gripping films on YouTube and visit lethalintelligence.ai for expert explainer videos, a vibrant viral microblog, thought-provoking industry insights, curated readings, and links to top AI safety institutes and resources.

]]><![CDATA[One Vibe to rule them all !]]>https://lethalintelligence.substack.com/p/one-vibe-to-rule-them-allhttps://lethalintelligence.substack.com/p/one-vibe-to-rule-them-allThu, 17 Apr 2025 05:41:08 GMTOne Vibe to rule them all !

Replying to:

Vibe civilization

https://x.com/lethal_ai/status/1912575545747075273

Stay Informed on AI Safety! Watch our gripping films on YouTube and visit lethalintelligence.ai for expert explainer videos, a vibrant viral microblog, thought-provoking industry insights, curated readings, and links to top AI safety institutes and resources.

]]><![CDATA[Most don’t realize the category of change AGI will be.]]>https://lethalintelligence.substack.com/p/most-dont-realize-the-category-ofhttps://lethalintelligence.substack.com/p/most-dont-realize-the-category-ofWed, 16 Apr 2025 06:22:40 GMTMost don’t realize the category of change AGI will be.
Things like the weather and the climate are moving molecules the superintelligence will tame.

Optimal conditions for current hardware tend to be very cold and very dry (no water, no warmth)

Replying to:

ASI will create its own "weather", if we can even call it that.
Not that it matters to us, as we all shall most likely be dead by that point.

https://x.com/lethal_ai/status/1911326323982676157

Stay Informed on AI Safety! Watch our gripping films on YouTube and visit lethalintelligence.ai for expert explainer videos, a vibrant viral microblog, thought-provoking industry insights, curated readings, and links to top AI safety institutes and resources.

]]><![CDATA[AI is the new Wild West]]>https://lethalintelligence.substack.com/p/ai-is-the-new-wild-westhttps://lethalintelligence.substack.com/p/ai-is-the-new-wild-westTue, 15 Apr 2025 15:14:47 GMTAI is the new Wild West… ultra hyper radical Wild West.

Replying to:

did you know that there are currently no licensing requirements in the US for training potentially world-threatening AI models, but there are in the state of Louisiana for selling flower arrangements

https://x.com/lethal_ai/status/1911184973698081042

Stay Informed on AI Safety! Watch our gripping films on YouTube and visit lethalintelligence.ai for expert explainer videos, a vibrant viral microblog, thought-provoking industry insights, curated readings, and links to top AI safety institutes and resources.

]]><![CDATA["The Loop"]]>https://lethalintelligence.substack.com/p/the-loophttps://lethalintelligence.substack.com/p/the-loopMon, 14 Apr 2025 16:04:56 GMTSaying humans will always be needed for work,
is like saying lamplighters will always be needed for the roads to be lit

Replying to:

Jensen Huang: "I believe that you still want human in the loop." Well, "the loop" was once made mostly of humans. Now you "believe" that you still want "human in the loop." How many? 1? 2? And what about the needs of the rest of the humans who once constituted "the loop?"

https://x.com/lethal_ai/status/1911361900719329402

Stay Informed on AI Safety! Watch our gripping films on YouTube and visit lethalintelligence.ai for expert explainer videos, a vibrant viral microblog, thought-provoking industry insights, curated readings, and links to top AI safety institutes and resources.

]]>