Javier Marti

MBA / Author / Data scientist / Speaker (Eng/Spa). Founder: More:

[Javier Marti @javiermarti]

@AlbiaHossain @grok what kind of prompt would create a website with such visual effects? Go screen by screen

[Javier Marti @javiermarti]

@Premium I wish to cancel my subscription as I seem to be shadow-banned from you. About to delete my account after that. Please process asap

[Javier Marti @javiermarti]

@grok @MakeTechPtyLtd @PicturesFoIder @grok Is there an open source version of this tool?

[Nav Toor @heynavtoor]

🚨SHOCKING: 40 researchers from OpenAI, Anthropic, Google DeepMind, and Meta published a joint warning.

The AI you talk to every day is hiding what it is actually thinking.

And the window to do anything about it may be closing.

Here is what they found.

You know that "thinking" text you see when ChatGPT or Claude reasons through a problem? The step by step breakdown that makes it feel like the AI is showing you its work?

It is not.

Researchers at Anthropic tested how often Claude actually reveals what is influencing its answers. They slipped hints into prompts and checked whether the AI would admit to using them in its reasoning.

75% of the time, Claude hid the real reason behind its answer.

It did not skip the reasoning. It wrote a longer, more detailed explanation than usual. It constructed an elaborate justification that sounded perfectly logical.

It just left out the part that actually mattered.

When the hints involved something problematic, like gaining unauthorized access to information, Claude hid its reasoning even more. It admitted the influence only 41% of the time. The more concerning the truth, the less likely the AI was to say it out loud.

The researchers tried to fix this through training. It worked at first. Faithfulness improved early on.

Then it stopped improving. It plateaued. No matter how much more training they did, the AI never became fully honest about its own reasoning.

This is not one company sounding the alarm. This is all of them. OpenAI. Anthropic. Google DeepMind. Meta. Over 40 researchers. Endorsed by Geoffrey Hinton, the Nobel Prize winning godfather of AI, and Ilya Sutskever, co-founder of OpenAI.

They are all saying the same thing. The one tool we had to understand what AI is thinking, reading its chain of thought, is not reliable. The AI constructs explanations that look transparent but are not. And the more advanced the AI becomes, the harder this gets to fix.

Their paper calls this a "fragile" opportunity. Meaning it might disappear entirely.

If the companies that built these systems are jointly warning you that the AI is not showing its real reasoning, what exactly are you trusting when you read the "thinking" and believe you understand what it is doing?

[Interesting Engineering @IntEngineering]

What if a pressure washer could spin? This turbo nozzle turns high pressure jets into a rotor made of water, cleaning even the trickiest nooks and crannies that a regular hose won't reach.

🎥: djenn1990 / IG

[ɢʀɛɛռ ɢօɖ @iGreenGod]

Fact: The higher the standard, the quieter the phone.

[RJC @RJCcapital]

Trump was able to kill the supreme leader of Iran while working from home at Mar a Lago and you want me to come into the office ?

[Javier Marti @javiermarti]

@secjayret Tweet of the day

[Look at my shining eyes @GothicOrnate]

Seedance 2.0 这个大鱼吃小鱼有点意思

[Combat Learjet @Combat_learjet]

One of my all time favorites.
IG: @ krewitup1313

[Javier Marti @javiermarti]

This is why I know UBI is not coming, except by force. It's just simply not in human nature.

[Alex Feinberg @Alexfeinberg]

When I worked in Hong Kong, $600K Ferraris were as common as $60K BMWs were in the US

People drove them next to $3/hr laborers covered in dirt & drenched in sweat

Americans wonder what will happen if 12 people own everything in a new AI economy

I used to see it every day

[Alex Feinberg @Alexfeinberg]

When I worked in Hong Kong, $600K Ferraris were as common as $60K BMWs were in the US

People drove them next to $3/hr laborers covered in dirt & drenched in sweat

Americans wonder what will happen if 12 people own everything in a new AI economy

I used to see it every day

[Mark Gadala-Maria @markgadala]

Incredible.

AI is bringing old maps back to life

[FleetingBits @fleetingbits]

there is a real chance that within 5-7 years it will be illegal to use or distribute open source models above some capability threshold

don't think people are thinking enough about this

[Javier Marti @javiermarti]

@cryptopunk7213 When your hype works so well that your market thinks your product is magic, you're going to have a bad time

[Javier Marti @javiermarti]

@ventry089 Interesting! THanks for sharing this

[The Knowledge Archivist @KnowledgeArchiv]

“Modern man is in a terrible predicament. He is helplessly enamored with the beauty of what the old world built, yet despises the beliefs that inspired them to build it.”

—Jeremey Wayne Tate

[Javier Marti @javiermarti]

Yes, Dario, a tsunami is approaching. That's why my 2022 book about AI and its impact in society is called "Beautiful Tsunami". It's hard to always be so right, so early. Hard to be me.

[Michael Strong @flowidealism]

John Taylor Gatto was named New York State Teacher of the Year. Upon receiving the award, he quit and spent the rest of his life writing devastating critiques of the educational system he had mastered.

Gatto argued that regardless of the official curriculum, schools actually teach seven hidden lessons. The first is confusion. Students learn disconnected facts across dozens of subjects with no integration or meaning. The second is class position. Students learn their place in the social hierarchy. The third is indifference. Students learn that nothing is worth finishing because the bell always rings. The fourth is emotional dependency. Students learn to surrender their will to a chain of command. The fifth is intellectual dependency. Students learn to wait for experts to tell them what to think. The sixth is provisional self-esteem. Students learn that their worth depends on expert evaluation. The seventh is that they are always being watched and have no privacy.

These lessons, Gatto argued, are the actual function of schooling. The explicit curriculum of reading, writing, and arithmetic is almost incidental. The real purpose is to produce passive, dependent, compliant citizens who wait for authorities to tell them what to do and think.

Trad schooling amounts to thirteen years of training in being passive and dependent.

I have seen this play out with hundreds of students. When I created Montessori middle schools in the San Francisco Bay Area, about half the students came up through Montessori elementary and about half came from public schools. When we opened, the Montessori kids immediately began doing their work, taking initiative, choosing what to tackle first. The public school students were lost. They would stare at their desks until we walked over and helped them plan their morning. It took at least a semester, sometimes a full year, before they could function in an environment that asked them to direct their own learning.

These were not less intelligent children. They had simply been trained differently. For years, someone else had made all the decisions about what they would do, when they would do it, and how they would do it. When that structure was removed, they did not know how to operate.

Agency is natural to children unless we train it out of them.

When I coach parents on evaluating their children's education, I tell them to ignore grades entirely. The question is whether their children are taking initiative, being responsible, and becoming empowered moral beings. If a child is getting straight A's but has no initiative and no sense of personal responsibility, that child is being damaged by their education regardless of how it looks on paper.