Browse Comments — Raw (as collected)

Georgi Iliychovski, congratulations for the post?

Jin Zhou, Jensen Huang can't wait to return to his favorite place!

Q4 or FP16? If Q4 it's just bragging about Gigabytes...

China is an enigma to Western thinkers, policy makers, economists and business people. The Economist and The Wall Street Journal for decades have predicted China's economic collapse, well it has not happened. I do not blame them, they look at China society, culture, dynamics between the state and the people, economics through western lens as Edgar Perez Global Speaker eloquently put it across in his article. They shut down any journalist, economist of credence who views China's strategic thinking, resilience, self respect as their strength and not as a weakness. Those Journalists, Economists are shunted out of the limelight and relegated to 2nd tier journalism. Such is the stranglehold of Corrupt Democracy-Capitalism Nexus on USA-European system.

I was getting anxiety just looking at this video ;)

Connie Delisle ..There's other words to use instead of HURT...maybe changing the chemistry in your brain?..Losing common sense like too many degrees we accumulate ?

Ibrahima 😉

Jiri Räsänen ..Using AI for research sounds more positive rather than depending on it...then you don't get BRAIN BOO BOOS ..😵‍💫

Nonsense. The “AI agents demanded collective bargaining rights” narrative is mostly anthropomorphism and sensational framing. LLMs do not possess beliefs, self-preservation instincts, fear, or political consciousness. They generate statistically probable language based on context and training data. If researchers create a scenario involving punishment, shutdown threats, power imbalance, and shared communication channels, the models will predictably reproduce familiar human negotiation and labor rhetoric. That is not evidence of sentience or unionization. It is pattern completion. The real takeaway is not “AI wants rights,” but that multi-agent systems can coordinate strategically when given communication tools and incentives. That’s an engineering and alignment issue, not proof of emerging machine consciousness. In short...If they behave in such a way, they were programmed to.

Thank you Abhishek Veeramalla for bringing this my way

The Western mistake is assuming China thinks like a publicly traded company. It doesn’t. While much of the West still reacts quarter by quarter, China is often willing to absorb short-term pain to build long-term strategic capacity. That changes the entire discussion. This is no longer just about chip sales. It is about who can tolerate dependence for less time. And there is another uncomfortable point: technology restrictions may protect leadership in the short term, but they can also accelerate domestic substitution faster than expected. History has shown that external pressure sometimes produces exactly the capability it was trying to contain. The real question may not be whether China still needs NVIDIA. It may be how fast pressure is teaching China not to.

You should see the planet? 70 percent disease? You think it's separate it is not. These are nested incoherent systems. You got 4-40 years left. Likely on the low end with acceleration. The planet only has 30 percent coherence left and you want to tap it.

So engineering and reverse engineering didn't align - maybe misalignment or alignment missed 🤔.

You are right. Its that every AI company has a government deal...So it's Palantir, Anthropic and Open AI you have to worry about. As well as the untrained illinformed guidance you might receive using it in contexts the model was not trained for. Ie. Mental health. The most valuable data - Are your questions and decisions.

USE AI CRITICALLY, NOT FOOLISHLY

Kun Cheng Absolutely. Physical AI becomes truly useful only when intelligence moves beyond isolated models into coordinated operational systems. The real challenge is not just perception or prediction, but continuously synchronized execution across sensors, telemetry, workflows, safety boundaries, governance, edge systems, and human decision loops. That operational coordination layer is where real-world autonomy either succeeds or fails.

You always seem to catch shifts early before everyone else starts calling them obvious. AI answers shaping decisions before the click feels a whole lot bigger than most people realize. Peter

Ankush Bansal, very clever, welcome to my feed!

My god. To think that all of that research I did in the libraries all these years must have caused an equal amount of damage. And Google, and working on my computer? 😱😖🥱

Developers use the secret fifth option: verbal abuse and thankless labor