Mohammed Daba @m_dabaAI

MS Robotics @UESTC1956 | Enjoying AI & Robotics linkedin.com/in/mohammed-da… Chengdu, China Joined June 2013

Tweets

302
Followers

37
Following

611
Likes

826

Mohammed Daba @m_dabaAI

2 weeks ago

@FishAudio DRAMA

0 0 0 13 0

View Details

@YiMaTweets “Frontier” alone doesn’t justify a course. But controlled experiments and predictions on unseen outcomes do. Students need to learn how to investigate unresolved questions before those questions are resolved. Otherwise, where would the people who resolve them come from?

0 0 0 208 0

View Details

Mohammed Daba @m_dabaAI

2 weeks ago

@MrEwanMorrison Aircraft are impossible to achieve true flight because they do not have feathers, biological muscles, or the evolutionary foundation that created birds, not a good argument

0 0 0 8 0

View Details

Mohammed Daba @m_dabaAI

2 weeks ago

Humanity spent millennia trying ways to organize discovery: universities, laboratories, scientific methods. AI could experiment with the rules of discovery themselves. What if the next breakthrough isn't a better mind, but a better way for minds to evolve together? (5/5)

0 0 0 7 0

View Details

Mohammed Daba @m_dabaAI

2 weeks ago

Humanity spent ~300,000 years as Homo sapiens, yet most of our technological acceleration came much later. Did our brains suddenly get smarter? Probably not. We changed the conditions under which intelligence could build on itself. What does that suggest about AI? (1/5)

4 0 0 24 0

View Details

Mohammed Daba @m_dabaAI

2 weeks ago

Here's a paradox: humanity's isolation may have helped innovation. Different communities explored different paths before exchanging. AI can communicate almost instantly, but could that make everyone converge same mistakes? Perhaps progress needs deliberate independence. (4/5)

0 0 0 12 0

View Details

Mohammed Daba @m_dabaAI

2 weeks ago

But copying human civilization would be a mistake. Evolution didn't optimize for efficient discovery. And 1,000 copies of one LLM aren't 1,000 independent minds: they may share blind spots. AI could deliberately create diversity in methods, models and representations. (3/5)

0 0 0 4 0

View Details

Mohammed Daba @m_dabaAI

2 weeks ago

Farming, writing and science weren't just inventions. They changed what future humans could discover. Maybe AI's next leap won't be a smarter model, but systems that build their own tools, experiments and environments, making previously impossible questions tractable. (2/5)

0 0 0 17 0

View Details

Mohammed Daba @m_dabaAI

2 weeks ago

@ChrisGPT The idea that now GTA development by AI is even discussed tell another story!

0 0 0 44 0

View Details

Mohammed Daba @m_dabaAI

2 weeks ago

@RohanArun no barriers of entry! How much money did you burn for the tokens?

0 0 0 27 0

View Details

Mohammed Daba @m_dabaAI

2 weeks ago

Much AI fear comes from seeing evolution as competition only. But life’s biggest leaps came from symbiosis: ancient bacteria cooperate with cells to became mitochondria, enabling complex life. AI with human may be similar. The future may be cooperation, not extinction.

0 0 0 16 0

View Details

Mohammed Daba @m_dabaAI

2 weeks ago

@fchollet Even if difficulty eventually matches each new AI capability level, couldn’t the transition, when stronger AI tackles the old difficulty frontier, as OA recent math results, produce a burstwith RSI delivering another cap jump, creating overlapping bursts (intelligence explosion)?

0 0 0 38 0

View Details

Mohammed Daba @m_dabaAI

2 weeks ago

@fchollet A much steeper difficulty curve that match RSI rate of development could look like an “explosion” on a human time scale, even though it remains mathematically linear.

0 0 0 38 0

View Details

Mohammed Daba @m_dabaAI

2 weeks ago

@fchollet You main argument center around the assumption that the increasing rate of difficulty will match AI development rate (so no ai explosion), but does history justifie applying that assumption to every possible form of AI self-improvement rate?

0 0 0 159 0

View Details

Mohammed Daba @m_dabaAI

3 weeks ago

emergent propety, hope it is clear who is right.

0 0 0 15 0

View Details

Mohammed Daba @m_dabaAI

3 weeks ago

Hard Problem of Wetness: 1. H₂O molecules aren’t wet 2. water is made of H₂O molecules 3. water is wet 4. so at least one of 1-3 must be false -Panpsychists say #1 is false -Dualists say #2 is false -Illusionists say #3 is false - Materialists say #4 is false

🐾Eneasz Brodski @EneaszWrites

3 weeks ago

the "Hard Problem" in a nutshell: 1. atoms aren't conscious 2. humans are made of atoms 3. humans are conscious 4. so at least one of 1-3 must be false Panpsychists say #1 is false Dualists say #2 is false Illusionists say #3 is false Materialists say #4 is false

1K 564 9K 2.0M 4K

1 0 0 38 0

View Details

Mohammed Daba @m_dabaAI

9 months ago

We are effectively the biological bootloader for silicon intelligence. At this point, the only thing we can hope for is that the bootloader remains compatible, and part of the future, once the system fully boots up.

0 0 0 36 0

View Details

Mohammed Daba @m_dabaAI

9 months ago

It’s a classic adversarial system (Moloch). Trust is impossible. Even if there were no geopolitical rivals like China, or if we signed a global treaty, the game theory remains: someone will always have the urge to progress in secret. If you pause, you lose.

1 0 0 39 0

View Details

Mohammed Daba @m_dabaAI

9 months ago

I think we’ve already passed the event horizon for AI superintelligence. We are being pulled toward the singularity, and just like falling into a black hole, there is no return path. You can’t fight physics (speed of light), and you can’t fight this momentum.