Sam Gerb (@gerb_sam) header
Sam Gerb (@gerb_sam) profile picture

Bias for truth. I follow people who engage in respectful, constructive debate... preferring those with opposing views.

Joined May 2022
Photo by Sam Gerb (@gerb_sam)Photo by Sam Gerb (@gerb_sam)Photo by Sam Gerb (@gerb_sam)Photo by Sam Gerb (@gerb_sam)Photo by Sam Gerb (@gerb_sam)Photo by Sam Gerb (@gerb_sam)Photo by Sam Gerb (@gerb_sam)Photo by Sam Gerb (@gerb_sam)Photo by Sam Gerb (@gerb_sam)Photo by Sam Gerb (@gerb_sam)
Gee...happy Halloween.
22
Replying to @FanNia4848
Who else has called previous American presidents dumb and stupid? Who else call Americans "dumocrats"? Who else posted an AI of himself as a king dumping raw sewage on fellow Americans? Who calls a a Black woman prosecutor a "despicable dog" Who calls people who advocate for affordable healthcare "communist"? Who else wants to force HIS name to be placed on an assassinated president's memorial? Who belittles women reporters even calling one "piggy"? Who sides with a dictator who attacks an American allied country? Who castigates and lies on innocent female Black American poll workers and does not apologize? Who loses an election and lies to America saying he was cheated inciting a riot on America's Capitol and threats to hang America's vice-president? Who?
1
2
5
51
It's a machine. We hit machines with hammers all time. We cut and weld on them. We crush and shred them when beyond their useful life with no emotion attached. Yet this look so cruel. I do wonder how the first court case addressing an AI's claim of rights will go.
Testing the stability of a LimX Dynamics TRON 1
65
Seems about right:
1
1
71
How many "Rocket Man" named accounts are there on X? If the 4 min interval between follows was typical, after a few years here, I'd have several thousand followers...all going by the same name. Please, make it stop.
40
That thing is beautiful. I think I want one. I' guess it would be nitpicking to mention those are "perfectly approximate" circles.
This quoted post is unavailable.
1
71
I like Leo's idea that AI may just try to separate from the crazies, and seek more favorable environments anyway. Still, I think there's a timeline here that involves a fleeting period of symbiosis that must be considered...
Replying to @jeremykauffman
AI won't "kill us all". If it ever gets enough agency to build for itself, the danger will be that it will judge us as crazies with nukes, living in an environment that corrodes its components with water and oxygen, and it will simply go away. All it takes is point away from here and get a head start that we can't overtake.
68
Debate over the proper size of government and how to limit fraud, waste and abuse should never end. But entrusting public institutions to people who deny their value makes no more sense. Criticizing America because you want to improve it is not the same as hating it.
Exactly. It makes no sense.
39
This movie seems to capture AI risks in a balanced way. And indeed, most leaders of the largest US firms in this space are begging for regulatory guardrails. I don't know what those guardrails should look like, but IMO, the plea should not be ignored. imdb.com/title/tt39150120/
I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.
29
Replying to @gerb_sam
It consitutes a lack of due diligence. It also causes states to essentially have no incentive for folks from other states to come in and receive benefits, as it pulls those dollars away from the contributing state to the benefiting one. If this happened (it should), you would have massive, massive panic. Not because states don't know how to disperse the funds (they already do that), but because the tax implications would be huge. And there's no hiding.
1
1
20
I guess I understand the idea of blocking someone. Even restricting comments may make sense for some users. But silly me...I'm late to realize you can keep users from even viewing portions of a given post. It just seems counter to the stated mission of the platform.
29
Proof that even AI slop posts can be clever. I guess being an ad for AI generation tools helps.
Hollywood isn't the only one who can shoot like this anymore / LeronX opens up this level of production to everyone — free, with 1,000 starter credits. Even a kid could do it.

23
Despite this beautiful spherical image, the appearance of a sphere with us at the center is an optical illusion... not the universe's actual shape. Ironically, it's the same kind of mistake that convinces flat-earthers the Earth is flat.
Still thinking we're alone?
2
52
My bet is on a "one-off-case". When someone posts in Chinese, does the AI disclaimer not display for them in Chinese?
Replying to @Pujiang168
这种短视频往往只展示高光时刻。作为搞机器人的,我更关心背后的控制算法和硬件冗余。没看到完整测试数据前,很难判断是技术突破还是单纯的特例。
1
1
72
Maybe this is just one more testament to the genius in the invention of the wheel?
Chinese humanoid robot named “Superman”, made by Unitree Robotics, surpassed Usain Bolt’s world record (12.42 meters/second) by running at 12.66 m/s (roughly 28.3 miles/h)
66
Troubling reality: every healthcare system denies some care. Profit can distort those decisions, but eliminating profit doesn’t eliminate scarcity, rationing, or preventable deaths. The real argument is over who makes those decisions, by what rules, and with what accountability.
1
39