Interesting 9 minute video snippet of Linus speaking with Dirk Hohndel, and giving his thoughts on Rust.

Including the video link would have helped: https://www.youtube.com/watch?v=0bb3-bjgf88

A very informative video snippet: Linus speaks with Dirk Hohndel about Rust in the kernel. Two key takeaways:

1. C is a simpler language ("it's one of the reasons I [Linus] enjoys C"), Rust is more complex. On the other side, Torvalds goes on to say, "Because it's simple, also very easy to make mistakes."

2. "If you're doing operating systems, you really don't have very many choices of languages: C, C-like language, or Rust.

You've got a huge, heavily tested code base that's decades old, full of myriad edge cases and weird cruft but it works correctly. It's built in an old language that starts with a "C".

Changing it correctly would require documenting all the requirements, building 20 million test cases, and require people skilled in the existing language and the new language, and they need to be highly conversant in both. And dealing with unexpected hurdles that couldn't be foreseen by the world's top programmers, like the maintainer describes (If Linus didn't see it, or other maintainers, it couldn't be seen until you got right on top of it).

You can call what keeps it going inertia, but everything has inertia, and that's not necessarily a bad thing.

"Few top executives have been more vocal in making the case for working from the office than Jamie Dimon, the veteran CEO of JPMorgan, who – as early as 2021 – sought to restore pre-pandemic working habits. “And everyone is going to be happy with it,” he told a Wall Street Journal event that year. “And yes, the commute – you know, people don’t like commuting. But so what?” - https://www.theguardian.com/bu...

I guess... nothing.

They do pay your wage. JP Morgan has been the beneficiary of great government largesse. And it took some convincing for labor to get a better deal from business owners.

They do pay you, and thus get to dictate a lot. But they don't own you. And labor can try to get a better deal.

Is not a threat to me or anyone I know.

Anybody who can just walk up to somebody they've never met and shoot them in the back is probably not the kind of person you want loose on the street anywhere.

Special operations commandos would like a word.

I think Luigi has the same internal controls as a commando. This wasn't a robbery, or for other personal gain, like a simple criminal. This was violence for a greater mission.

The only way to fix housing is to build housing.

It's not clear that the cause of this is a lack of housing. There are as many housing units per individual and per household as there were at the height of the first housing bubble, the last maximum. If the problem is too many investors with too many housing units, you can't build your way out of that.

Now... if they're undercounting immigrants, of which we had a surge, both legal and illegal, then perhaps that wouldn't show up in the official number of housing units per household and per person.

The question is not the 88 dollars, it's the precedent it sets for future collusion and price increases. It won't stay at 88 dollars if they can get away with it.

Question the collusion, not the result of the collusion. The collusion results in landlords having too much power, like any monopoly. The software enables the de facto monopoly.

"If they can get you asking the wrong questions, they don't have to worry about the answer." - Thomas Pynchon.

This is why free market fundamentalism (the belief that a free market unregulated by government always generates socially optimal outcomes) doesn't work: companies don't only compete in the marketplace but in advertising and law. The response might be, "if only there were no law applying to business, it would lead to socially optimal outcomes" - well, then mafias could thrive (for example), and businesses (collections of people) would be free to be lawless.

So - laws must apply to business. And business will compete in the legal arena to obtain maximum advantage.

Something happened in the aughts, the years 2000-2009. The various financial bubbles sparked or peaked, medical, education, housing. Personally that's when I noticed the quality of durable goods declining dramatically. A new social paradigm took hold, driven by companies seeking to extract maximum value without regard to longer term company performance or product quality. Income inequality grew. There was some sort of paradigm shift, I suspect in business schools and in monetary policy, which has been wealth-extractive to the majority of society.
 

I know workaholics. They have been blessed/cursed to find a task which engages and consumes them fully. They are typically highly successful, in whatever field they are in. Their vocation is their avocation.

What they fail to realize is the vast majority are not, nor will ever be, workaholics. People can be very energetic, but on hobbies (their avocation), not their work (vocation).

Considering they require the cooperation of the masses to realize their vision, and the masses do in fact, feed and protect them (Indira Gandhi was assassinated by her bodyguards), they should take into account that not everyone's vocation is their avocation.

How are you different than a database? What is your brain doing that a database can't?

The first big difference between an LLM and a human is that the LLM is set in stone after training. It is a static filter. It gains no more information. The structure is set. With a human, it can immediately update its own personal knowledge store (database) with new information. It's also able to autonomously decide to do this. "Berries taste good. Red berry taste good, make Gorok feel good. Blue berry make Gorok feel bad." Database updated. After training, the LLM never updates with "blue berry bad". It would require another multi-week, round of training to re-set the weights.

Are we next token generators? It would not surprise me if we are something like it. My understanding is (I'm looking for the CACM article but cannot find it right now) that the token response is built by running the response token through the neural network serially and repeatedly (the initial training of the neural network is done with a parallel process using GPU's). We appear to be massively parallel. We could be doing next token generation in a ultra low power, massively parallel manner.

A part of our brain is doing what a database does. I think the information retrieval piece is probably hugely more powerful in an LLM due its massive database being highly optimized to infer meaning and retrieve information.

We also have the ability to navigate effortlessly in the real world, after training in infancy. Developing a representation of the physical world is currently outside the reach of the LLM but I suspect that's a matter of time before it can. It's power requirements will likely be monstrous, and we are ultra lower power (they're talking about harnessing nuclear reactors to power AI / LLM operations).

I certainly don't think we are utterly different. The whole concept of the neural net is based loosely on the brain's design. Nature has already solved a lot of problems to which humans are only catching up, still trying to mimic nature in many ways, versus improving on it. The neural network theory has been around for decades, but breakthroughs in both software, computation, and hardware has enabled this current functionality, it seems to me.

Again, I'm not an expert so check my work and YMMV.

An AI / LLM is a database you can talk to (query) using natural language. It is an amazing achievement, famously fooling one of Google's own software engineers (Blake Lemoine) into believing the machinery was sentient.

But it's still a database. It's trained for weeks on a datasets to set the weights in the neural network. Tokens from prompts filter through the (static) neural network repeatedly building a response ("inferring").

The problem is as the tokens cycle through the neural network, building the response by filtering through the weights, it's impossible for a human to know exactly what it's doing - specifically how it's reasoning to come to its conclusions. That's where the field of Explainable AI comes in.

To help people get a handle on AI, here's how they're priced - based on tokens:
https://learn.microsoft.com/en...
https://help.openai.com/en/art...

A bunch of weights in a neural network (the weights set by weeks of continuous training on a dataset), tokens extracted from prompts, filter through the neural network, building the output. Is the possibility of sentience in there? Consciousness? What's the core action being taken? How exactly is a token response built?

Here's a description from IBM:

During the training process, these models learn to predict the next word in a sentence based on the context provided by the preceding words. The model does this through attributing a probability score to the recurrence of words that have been tokenized— broken down into smaller sequences of characters. These tokens are then transformed into embeddings, which are numeric representations of this context.

To ensure accuracy, this process involves training the LLM on a massive corpora of text (in the billions of pages), allowing it to learn grammar, semantics and conceptual relationships through zero-shot and self-supervised learning. Once trained on this training data, LLMs can generate text by autonomously predicting the next word based on the input they receive, and drawing on the patterns and knowledge they've acquired. The result is coherent and contextually relevant language generation that can be harnessed for a wide range of NLU and content generation tasks.

Once it's trained, it's trained. The neural network and its weights are set. Then it's time to query / prompt.

It's amazing stuff. That the machinery can do this is astonishing. But it's feeding tokens through a trained, static neural network.

Now... right now it's trained on binary data, audio, video, images, text. Prompts are tokenized from incoming text strings. The technology is in its infancy. Could you program something around this core system to be a decision making platform that could be placed in a robot which could navigate its environment, and make decisions about what to do? I think that's coming. That would require being able to tokenize the world around it. I suspect it would take a vast amount of training data that may be beyond current computing and electrical power capabilities. This is nascent technology and there's a long road ahead.

On the other hand... quantum computing, fusion... these have promise, but engineering limitations limit the ability to realize those promises. So, one needs to have a balanced view. We're just at the beginning though. Relational databases were introduced in the early-mid 70s. This technology has been introduced just now, so who knows what it'll look like in 50 years.

Disclaimer: I'm not remotely an LLM / AI expert. Reading the CACM, thinking about it, but there are lots of people out there programming these things, of which I'm not one. But I am interested and think I have it right.

As this type of study would be both illegal and impossible to implement for this investigation, other sources of data must be used to draw conclusions. Shaken Baby Syndrome was pieced together from the evidence of a hundred years of observation (see PDF), noting the frequency and type of brain injury that appeared with abuse-related trauma. In the early 70s, the syndrome was finally described, and per the links I posted above from neurosurgeons' and pediatricians' associations, they have concluded the syndrome exists.

These are two organizations representing experts in the field of both neurosurgery and pediatrics, and don't appear to have ulterior motives, so their conclusions are almost certainly accurate.

Those denying the syndrome do have ulterior motives, namely, freeing their client by casting doubt on the very existence of the syndrome.

1) Statement from American Association of Neurological Surgeons (AANS): "Shaken Baby Syndrome (also known as Shaken Impact Syndrome) is a serious form of abuse inflicted upon a child. It usually occurs when a parent or other caregiver shakes a baby out of anger or frustration, often because the baby will not stop crying."

The AANS appears to be a legitimate association for neurological surgeons. Google this phrase to see how it is referenced by colleges and universities: american association of neurological surgeons site:*.edu

2) The Scientific American article does not deny that Shaken Baby Syndrome exists. It says, "There is no doubt that shaking a child can cause injuries, including those that comprise the shaken baby syndrome triad." It however goes on to advocate for Roberson, saying, "Newer research, however, has shown that shaking is not the only way to cause those injuries". Newer research? I'm sure there have been multiple ways to cause those injuries for a long time.

3) A doctor's blog on Shaken Baby Syndrome : "As a pediatric intensivist for over 30 years, I have dealt with many unfortunate examples of this entity, and I have no doubt that it exists. But, like all disorders that do not have a specific, definitive test for them, deciding whether or not a child has suffered shaken baby syndrome depends upon more that some x-rays and an eye examination; you need to consider the entire context of the story."

4) The American Association of Pediatrics statement on the WaPo article (linked in the blog article above - appears to be paywalled): "What are the facts? [about denying the existence of the syndrome] it involves a tiny cadre of physicians. These few physicians testify regularly for the defense in criminal trials — even when the medical evidence indicating abuse is overwhelming. They deny what science in this field has well-established. They are well beyond the bounds where professionals may disagree reasonably. Instead, they concoct different and changing theories, ones not based on medical evidence and scientific principles. All they need to do in the courtroom is to obfuscate the science and sow doubt."

Remember the insane covid doctors? Yeah, me too.

Wiki says AAP is the "largest professional association of pediatricians in the US."
____________________________________
An interesting phenomenon happens when the anti-death penalty movement makes a full court press for one convict. They are able to create a nationwide sensation for the convict and cast doubt on whatever is necessary to create doubt about the offender's conviction. It's a fascinating thing to watch, part of that "fake news" phenomenon, i.e. disinformation on a large scale, through existing channels.

In this case, they're claiming Shaken Baby Syndrome is junk science. They bring out doctors who say that's the case. Media is amplifying the disinformation. The truth is lost, as it is supposed to be in this type of situation.

A doctor quoted on the radio put it simply: "The only people who are claiming shaken baby syndrome is junk science, are outside of the medical profession."