Robert Wu

CFA, MBA

[Robert Wu @balanceview]

$tlt will do well for next year. A rebalancing toward fixed income is happening.

[Robert Wu @balanceview]

@JG_Nuke It is more meaningful how much AR grow as a percentage of sales.

[Robert Wu @balanceview]

Unpopular opinion: if you can solve energy storage problem (safety, efficiency, costs), nuclear power will be an inferior option.

[Robert Wu @balanceview]

If you still can't see that $tsla will win the self-driving battle handily, you are blinded by your ideology. The technology has advanced to the level that will force legislative action soon.

[Robert Wu @balanceview]

Can Human learn everything from language alone? Probably not. More computer power for LLM will lead to #AGI is questionable.

[Robert Wu @balanceview]

Lidar is not the way to go! #google wake up. https://t.co/Ci8Rq9Ptlh It is not too late to change course. At least to start an approach by using #AI and sensors only.

[Robert Wu @balanceview]

@JustinTrudeau What a stupid post. You could lose the game anytime. Is this the same way you felt about Canada?

[Robert Wu @balanceview]

calculate the most relevant ones. How to make that judgement is the key.

[Robert Wu @balanceview]

Human has common sense and we discard most of the variables when we make judgment. Today's AI is using all variables available to it and calculates the optimized outcome. Inevitably, it is getting more complicated when more variables are introduced and more computing powers

[Robert Wu @balanceview]

are needed. It is unsustainable in this form. More powerful chip, more energy consumptions and more variables will be introduced. That's why Deepseek is forcing people to think differently. Eventually, AI needs to be more like human to be able to discard most variables and only

[Robert Wu @balanceview]

@chh2os @squawksquare No, Lidar doesn't help much. two issues: 1. car can't follow lane, instead it should follow plowed route. It is common sense, should be easily learned by AI. 2. sensors maybe covered by ice and snow. This should be solved by an option to heat up sensor to 40F.

[Robert Wu @balanceview]

Don't try fsd in winter storms. Not working yet.

[Robert Wu @balanceview]

Would love to invest in Palmer Luckey's company. His smartness and determination is close to Elon's level.

[Robert Wu @balanceview]

$tsla's #FSD is not useful in heavy snow. I hope someone is working on it. One third of the world has winter storms.

[Robert Wu @balanceview]

@elonmusk Don't like to drive in heavy snow. Hope Fsd can solve the snow problem in next update. Currently it isn't useful in heavy snow

[Robert Wu @balanceview]

#FSD needs to solve snow problem. When snowstorms cover road marks, or traffic follows snow plowing paths instead of line marks, it's difficult for fsd. $TSLA needs to solve it before #cybercab becomes a reality

[Robert Wu @balanceview]

@SawyerMerritt #Lidar is not the way to do it. It becomes clearer everyday.

[Robert Wu @balanceview]

#GPU ($nvda) as a foundation of #AI is unsustainable. The huge power consumption will lead to either 1.much more nuclear power (nobody likes to live close to it) or 2. more solar power (battery storage will be huge which benefit $tsla well). But I will bet some other types of chips (eg. NPU) with much less power consumption will emerge.

[Robert Wu @balanceview]

Lidar approach for self-driving is too complex, expensive, ugly and difficult to scale. $tsla way is more nature and requires a lot of AI training (it is getting better every day due to the vast data). Most market share for self-driving will be $tsla licensed