Hacker News Reader

Comment by mlyle | Hacker News Reader

Nearly as bad: trying to use systems made out of meat, evolved from a unrelated background and trained on an undocumented and chaotic corpus of data, to try and produce content satisfying a strict mathematical model.

View on HN

Except that the "systems made out of meat" are the entities which both define the problem needing to be solved and are the sole determiners if said problem has been solved.

Of note too is that the same "systems made out of meat" have been producing content satisfying the strict mathematical model for decades and continue to do so beyond the capabilities of the aforementioned algorithms.

+1. Also, those systems made out of meat were the ones that discovered the strict mathematical models in the first place and the rules by which they work.

It's usually not the same pile of meat defining the problem and solving the problem.

Yes, humans exceed the capability of machines, until they don't. Machines exceed humans in more and more domains.

The style of argument you made about the nature of the machinery used applies just as well (maybe better) to humans. To get a valid argument, we'll need to be more nuanced.

>> Except that the "systems made out of meat" are the entities which both define the problem needing to be solved and are the sole determiners if said problem has been solved.

> It's usually not the same pile of meat defining the problem and solving the problem.

True, but this distinction is also irrelevant.

The point is that problems capable of being solved by software systems are identified, reified, and then determined to be solved by people. Regardless of the tooling used to do so and the number of people involved.

> Yes, humans exceed the capability of machines, until they don't. Machines exceed humans in more and more domains.

But machines do not, and cannot, exceed humans in the domain of "understanding what a human wants" because this type of understanding is intrinsic to people by definition. Machines can do a lot of things, things which can be amazing and are truly beneficial to mankind, but they cannot understand as people colloquially use this term since they are not people.

I believe a decent analogy for this situation is how people will never completely understand the communication whales use with each other the way whales do themselves. There may someday exist the ability to translate their communication into a semblance of human language, but that would be only what we think is correct and not the same as being a whale.

What if systems develop that are able to understand typical human intention and wants with an accuracy that is superior to most humans?

You seem to rule this out, but despite having similar biology and wants, humans misunderstand others' intents and miss cues a lot.

Is it impossible that humans could build a system to know what a whale wants, based on its vocalization, that does better than the typical whale? Do we know that whales do really great at this, even?

> Is it impossible that humans could build a system to know what a whale wants, based on its vocalization, that does better than the typical whale?

The answer resides in the question. For a human to know what a whale wants is to have an understanding of what it is to be a whale, not deriving a Rosetta Stone[0] such that bidirectional vocalizations could be exchanged.

So is it possible to communicate with a whale? Sure.

But the infeasiblity of cross-species understanding is far more difficult than even what you identified above:

  ... despite having similar biology and wants, humans 
  misunderstand others' intents and miss cues a lot.

Now add to all of this the fact that computer algorithms are just that, and not even having the shared commonality of being mammals, and the possibility of a machine understanding what a human wants is fantastical at best.

0 - https://en.wikipedia.org/wiki/Rosetta_Stone

I think this is a silly take. Not all species had evolutionary pressure for cross-organism understanding. A different species that has selective pressure for social understanding and greater mental understanding is probably going to understand what an insect wants better than say, another insect.

That is, how well you can understand someone else doesn't depend just on common ground, but perhaps even more on how mature of machinery for understanding that you have.

> That is, how well you can understand someone else doesn't depend just on common ground, but perhaps even more on how mature of machinery for understanding that you have.

I disagree, and believe I can explain why.

If I wrote:

  I can think of a few HN posts I have written which did not
  properly express my intent when I wrote them.

Is this a statement that you can relate with?

Even if not, does it communicate an experience you could envision based on your being a participant in this forum, having written posts yourself, and knowing I am a person just as you are doing the same?

What if I texted the sentence to someone who has never heard of HN?

Now what would happen if I wrote that same sentence on a piece of paper and handed it to a random person in a train station where no one spoke English?

The difference: Meatbags created something like an education system, where the track record in it functions as a ledger for hiring companies.

There is no such thing for AI. No ledger, no track record, no reproducibility.

Batter. Meatbags created liability system where if meatbag harms others they go to jail.

On the other hand, some of the most capable meat bags said fuck you to your record keeping system and dropped out.

Those meatbags usually have a different form of ledger. Be it open source contributions, built prototypes, apps or games, or other things they've shared before to create a track record of their skillset.

I never claimed that there is one type of ultrageneric ledger that works for all areas of research. But somehow, the LLM world still thinks that is the case for whatever reason.

Meaty feet can be held to a fire. To quote IBM, "A computer can never be held accountable."

This is the question I keep asking leaders (I literally asked a VP this question once in an all hands). How do we approach the risk associated mistakes made by AI?(process, legal, security, insurance etc) We have process and legal agreements in place to deal with humans that work for a business making mistakes. We need analogs for AI if we want to use it in similar ways.

What was the answer? Asking for a vp friend

>>> Meaty feet can be held to a fire. To quote IBM, "A computer can never be held accountable."

>> This is the question I keep asking leaders (I literally asked a VP this question once in an all hands). How do we approach the risk associated mistakes made by AI?

> What was the answer? Asking for a vp friend

This is a difficult issue to tackle, no doubt. What follows drifts into the philosophical realm by necessity.

Software exists to provide value to people. Malicious software qualifies as such due to the desires of the actors which produce same, but will no longer be considered here as this is not germane.

AI is an umbrella term for numerous algorithms having wide ranging problem domain applicability and often can approximate near-optimal solutions using significantly less resources than other approaches. But they are still algorithms, capable of only one thing - execute their defined logic.

Sometimes this logic can produce results similar to be what a person would in a similar situation. Sometimes the logic will produce wildly different results. Often there is significant value when the logic is used appropriately.

In all cases AI algorithms do not possess the concept of understanding. This includes derivatives of understanding such as:

  - empathy
  - integrity
  - morals
  - right
  - wrong

Which brings us back to part of the first quoted post:

  To quote IBM, "A computer can never be held accountable."

Accountability requires justification of actions taken or lack thereof, which demands the ability to explain why said actions were undertaken relative to other options, and implies a potential consequence be imposed by an authority.

Algorithms can partially "justify their output" via strategic logging, but that's about it.

Which is why "a computer can never be held accountable." Because it is a machine, executing the instructions ultimately initiated by one or more persons whom can be held accountable.

In the all hands I got an answer about techniques that would be used to reduce the likelihood of mistakes. Ie not an answer.

I suspect the analog will be that the "human in the loop" will bear all the consequences. Perhaps even if they did nothing wrong and are in fact the victim in that situation.

Take the case of Linda Yaccarino. Ordinarily, if a male employee publicly and sexually harassed his female CEO on Twitter, he would (and should) be fired immediately. When Grok did that though, it's the CEO who ended up quitting.

My question is if I get some code from AI, save it to a file, then modify it or add some functions to it, can I still claim the copyright for it at the top of the file? Do I need to give the AI any credit?

I'm asking because I read somewhere that "AI produced output cannot be copyrighted". But what if I modify that output myself? I am then a co-creator, right, and I think I should have a right to some copyright protection.

You do.

First, a few disclaimers, I am not a lawyer and this is an actively evolving area.

The answer that most aligns with current precedent to my knowledge is that the parts you modify are protected by your copyright, but the rest remains uncopyrightable. With the exception of any chunks generated that align with someone's existing copyrighted code, as long as those chunks are substantial and unique enough.

This is how I understand it as well. There was a case of a monkey taking a selfie and the photog attempting to assert copyright. Courts determined (if I remember correctly :p ) that only human works are copyrightable and therefore the monkey's selfie could not be copyrighted at all.

I deleted a whole paragraph with exactly that example, so at least if we're wrong, we're wrong together.

> "A computer can never be held accountable."

cf the Post Office scandal in the UK which was partly helped along by the 1999 change in law[1] which repealed the 1984 stance that "computer evidence is not permissible unless it is shown to be working correctly at the time"[0]; i.e. that a computer was now presumed to be working correctly and it was up to the defence to prove otherwise.

[0] https://www.legislation.gov.uk/ukpga/1984/60/section/69/1991...

[1] https://www.legislation.gov.uk/ukpga/1999/23/section/60/1999...

Countless bodies consisting of the said meat have been responsible for the advancement of technology so far. If these meat brains don't contribute to any new advancements then the corpus of data will stay stagnant!

Where do you think training data comes from

Those systems of meat were trained over hundreds of millions of years compared to mere months or years.

I'm getting so tired of this dumb kind of non-argument. You can't defend LLMs on their own merits, so you try to make them look smarter by throwing shade on humans. That's a non sequitur and whataboutism.

Nah-- I feel like I have my eyes pretty wide open about the shortcomings of LLMs (but still find them useful often).

But any argument seeking to dunk on LLMs needs to not also apply equally to the alternative (humans).

And, wouldn't you know it, it actually does not also apply equally to the alternative (humans).

How much of it doesn't? We're deterministic? (aren't we less deterministic than LLMs?) All of our training is auditable? (there's a wealth of unknown experiences in each person writing code, to say nothing of the unknown and irrelevant experiences in our evolutionary background).

Maybe you can argue we don't use statistical completion and prediction as a heavy underpinning to our reasoning, but that's hardly settled.

Nah-- you will have to try harder to make an argument that really focuses on how LLMs are different from the alternative.

I find this take to be purely misanthropic. we are more than stochastic parrots

Reply on news.ycombinator.com