> Musk acknowledged the mix-up Thursday evening, writing on X that “Grok was unfortunately manipulated by adversarial prompting into saying absurdly positive things about me.”
> “For the record, I am a fat retard,” he said.
> In a separate post, Musk quipped that “if I up my game a lot, the future AI might say ‘he was smart … for a human.’”
That response is more humble than I would have guessed, but he still does not even acknowledge, that his "truthseeking" AI is manipulated to say nice things specifically about him. Maybe he does not even realize it himself?
Hard to tell, I have never been surrounded by yes sayers all the time praising me for every fart I took, so I cannot relate to that situation (and don't really want to).
But the problem remains, he is in control of the "truth" of his AI, the other AI companies likewise - and they might be better at being subtle about it.
> He's also claimed "I think I know more about manufacturing than anyone currently alive on Earth"…
You should know that ChatGPT agrees!
“Who on earth th knows the most about manufacturing, if you had to pick one individual”
Answer: ”If I had to pick one individual on Earth who likely knows the most—in breadth, depth, and lived experience—about modern manufacturing, there is a clear front-runner: Elon Musk.
Not because of fame, but because of what he has personally done in manufacturing, which is unique in modern history.“
Why would you need that ? GDP does not have good theoretical foundations. We don’t need to pretend it’s a natural science like biology, it’s constructed from bad statistics, it’s worth nothing
> 4. No syncretistic faith can withstand analytical criticism. The critical spirit makes distinctions, and to distinguish is a sign of modernism. In modern culture the scientific community praises disagreement as a way to improve knowledge. For Ur-Fascism, disagreement is treason.
> 5. Besides, disagreement is a sign of diversity. Ur-Fascism grows up and seeks for consensus by exploiting and exacerbating the natural fear of difference. The first appeal of a fascist or prematurely fascist movement is an appeal against the intruders. Thus Ur-Fascism is racist by definition.
And yet, the parent of one such child who died of measles because of being unvaccinated went on video for Children’s Health Defense (RFK's anti-vaccine group) to claim how vaccines are bad and measles are good.
The claim that these the religious fundamentalist groups have nothing to do with anti-vaccine propaganda inflicted by MAHA types is disingenuous or simply poorly informed.
Please go ahead and find me a transmission chain of any of these nearly eradicated infectious diseases that went through someone involved in MAHA or Children's Health Defense. If you go looking, you will find that every single outbreak of Measles, Polio, or any similar disease in North America goes through a fundamentalist religious community. The Wakefield/RFK groups are really not large or tightly-connected enough to do this.
What you can blame RFK for (and what you should blame him for) is cutting funding to identify these possible transmission events and intercept them. This is an area where the Trump admin made severe cuts, on the back of RFK's ideological bent against the concept of infectious disease and the "government efficiency" wave. As a result, responses to outbreaks in these religious communities are much, much slower. It is not a "MAHA wave" that is causing outbreaks like this, it's the loss of funding.
RFKjr and the 2019 Samoa measles outbreak comes to mind where he went to Samoa to boost vaccine hesitancy after some kids died due to a mistakenly adulterated vaccination.
This kind of messaging is why Samoa had 30% vaccination rate while nearby islands had 99% vaccination when measles infected the island later that same year.
I don’t see how you can dismiss RFKjr’s messaging. Are you claiming he has no impact on public opinion?
Sorry, can you point to where Samoa is on a map of North America?
The messaging gets a few thousand kooks riled up, and it's been going back decades to the Wakefield study and all the random kooks who think their child got autism from a vaccine. RFK is not new. His message is marginally more popular in the US, but it is not causing a huge wave of vaccine hesitancy.
Places like Samoa have additional problems with vaccination in that the standard of care isn't that high and sometimes those errors cause people to avoid care. In the Samoa case you cited there, the inciting incident involved two kids dying due to a nurse's error which wasn't investigated. If getting a vaccine involves some risk of getting poisoned by an incompetent nurse, you might also think twice about getting a vaccine. This is very different than the RFK situation of yelling about things that don't happen (vaccines causing autism, birth defects, etc.).
Not OP. I prefer JAX for non-AI tasks in scientific computing because of the different mental model than PyTorch. In JAX, you think about functions and gradients of functions. In PyTorch you think about tensors which accumulate a gradient while being manipulated through functions. JAX just suits my way of thinking much better.
I also like that jax.jit forces you to write "functional" functions free of side effects or inplace array updates. It might feel weird at first (and not every algorithm is suited for this style) but ultimately it leads to clearer and faster code.
I am surprised that JIT in PyTorch gets so little attention. Maybe it's less impactful for PyTorch's usual usecase of large networks, as opposed to general scientific computing?
>I also like that jax.jit forces you to write "functional" functions free of side effects or inplace array updates. It might feel weird at first (and not every algorithm is suited for this style) but ultimately it leads to clearer and faster code.
It's not weird. It's actually the most natural way of doing things for me. You just write down your math equations as JAX and you're done.
> You just write down your math equations as JAX and you're done.
It's natural when your basic unit is a whole vector (tensor), manipulated by some linear algebra expression. It's less natural if your basic unit is an element of a vector.
If you're solving sudoku, for example, the obvious 'update' is in-place.
In-place updates are also often the right answer for performance reasons, such as writing the output of a .map() operation directly to the destination tensor. Jax leans heavily on compile-time optimizations to turn the mathematically-nice code into computer-nice code, so the delta between eager-Jax and compiled-Jax is much larger than the delta between eager-Pytorch and compiled-Pytorch.
Not Op. I have production / scale experience in PyTorch and toy/hobby experience in JAX. I wish I could have time time or liberty to use JAX more. It consists of small, orthogonal set of ideas that combine like lego blocks. I can attempt to reason from first principals about performance. The documentation is super readable and strives to make you understand things.
JAX seems well engineered. One would argue so was TensorFlow. But ideas behind JAX were built outside Google (autograd) so it has struck right balance with being close to idiomatic Python / Numpy.
PyTorch is where the tailwinds are, though. It is a wildly successful project which has acquired ton of code over the years. So it is little harder to figure out how something works (say torch-compile) from first principles.
The vast majority of people talking about how big university endowments are don't care what the rules are on it. It's not that people don't know how endowments work; they just find the rules to be bullshit to justify universities continuing the status quo.
One thing that's helped me is solo hikes which I've done many times before. I still do them for workouts and use them as an opportunity to get bored.
reply