When that happens typically the company starts optimizing for sucking money from the government. From the point of view of the consumer Intel would be finished.
Has anyone tried to see what the output looks like if the model is never allowed to be uncertain?
For example, whenever certainty drops below a threshold the sampler backtracks and chooses different tokens. Such that at the end every single token had an above threshold certainty.
I doubt it would entirely eliminate undesirable outputs, but it would be interesting.
Tragedy of the commons is why short term is all that matters and will ever matter to non-ideological investors.
If an action that hurts the stock short-term but will help int he long-term needs to be performed why would you as an investor enact it or even stay for the ride?
You are better off either opposing it or selling your stock and then waiting to see if someone will enact the changes, then you have the "insider" information to know that the short-term stock drop was a good thing for the long-term and rebuy the shares cheaper.
Since investors are forward-looking, I don't think there are many situations where a change that's good for the long-term value of the company will necessarily drive down the share price in the short term.
You might drive down the share price if earnings go down, but that's not necessarily the case if the long-term value is clear. An example would be any company reinvesting earnings in a new factory.
>If an action that hurts the stock short-term but will help int he long-term needs to be performed why would you as an investor enact it or even stay for the ride?
Yes and this underpins what most consider the disgusting trait of being excessively greedy, not just executives, doctors specialties like dermatology, anything.
The doctor who wants the ultra high paying medical specialty sometimes cares nothing about becoming rich. Fear. Fear of being crushed by all that student debt. Doing everything possible to avoid that. Planning to start a family, fear of homeless, etc. Fear of being unable to retire. Excess, irrational fear. Then, it passes, and now we have what seems in retrospect, simply a greedy bastard with now an excessive amount of money.
Same with executives, the thought may not be "let me cheese short term gains as hard as possible" but "let me hedge against short term losses as much as possible".
Aggression from sheer greed is human human thought modality. Predatory. Aggression when cornered, anxiety / fear response, arguably a more vicious and nasty aggression, is another very much different type of aggression because it is not predatory. The difference is predatory feels more voluntary and fear is about whether I should risk being super aggressive or not. In defense-of-self aggression, fear as already there plenty, so a person can do about any hyper aggressive evil thing in the name of defending from a perceived threat - such as lay off half the work force, destroy customer relationships, etc. In this way a good leader must have a steady hand and outlook with regards to fear.
> Tragedy of the commons is why short term is all that matters and will ever matter to non-ideological investors.
Tragedy of the commons was an ideological essay designed to justify privatization of public goods. It was disproven by data before it was published. I am sure there are some hyper specific examples where it has happened as described, but as a “fact” about the world and as a justification for any course of action, it’s highly suspect.
I don't disagree that it was an ideological "essay" but I dunno why you're linking that when it's most associated with Aristotle, in particular:
"What is common to many is taken least care of, for all men have greater regard for what is their own than for what they possess in common with others."
Maybe you're suggesting that that essay popularized the phrasing? But I'm pretty sure even as a coined "term" it was around before then
This is the first I've heard that it's false; however it seems like many times in my life I've observed something suffering, seemingly from lack of ownership despite being a common good.
What do you call that if you can't call it a tragedy of the commons?
On HN I’ve seen regular claims that “tragedy of the commons has been disproven”. I’ve not yet identified which social bubble propagates this or what it is based on but there seems to be some niche in which people are being taught that it is categorically proven to be an invalid concept.
That persons argumentation is awful, however. Another patient soul took the time to challenge the idea with logic, and there wasn’t much actually supporting the idea in response.
> I am sure there are some hyper specific examples where it has happened as described, but as a “fact” about the world and as a justification for any course of action, it’s highly suspect.
You read that and understood “categorically proven to be an invalid concept”?
People cite the tragedy of the commons to discourage sharing of resources. The idea that common land should be divvied up into private ownership to prevent them falling into ruin. When really shared resources just need accountability between the people who make use of them. It’s just basic game theory:
If there is no cost to abusing your opponent, that strategy will get used. But if you’re going to be playing long term with the same people, systems will form to deter abuse.
Tragedy of the common is a concept dating back to ancient times that has extremely broad empirical support throughout history. The essay you mention took the name from all this real world experience, and even if the essay is bad, the concept is anything but.
A simple search finds more examples and references to literature than you can likely read in years.
Dating back to ancient times - specifically Aristotle. Aristotle also believed that women are deformed men. That some people deserve to be enslaved. That deaf people are incapable of reason.
So, just because it’s old doesn’t mean it’s true. Just like “kids these days” or “seems like nobody wants to work anymore” some bad ideas are evergreen.
When the UK enclosed their fields so that they were privately controlled crop yield immediately went up 20%. Theres overwhelming empiric evidence of tragedy of the commons being real.
it’s a specific viewpoint that chooses to measure success in terms of crop yields, rather than quality of life…
The tragedy of the commons didn’t claim that crop yields are slightly depressed. The claim is that the commons will be overused to the point of threatening to destroy it.
The race for true AI is on and the fruits are the economic marginalization of humanity. No game theoretic actor in the running will shy away from the race. Anyone who claims they will for the 'good of humanity' is lying or was never a contender.
This is information technology we are talking about, it's virtually the exact opposite of nuclear weapons. Refining uranium vs. manufacturing multi purpose silicon and guzzling electricity. Achieving deterrence vs. gaining immeasurable wealth, power and freedom from labor.
This race may even be the answer to the Fermi paradox - that there are few individual winners and that they pull up the ladders behind them.
This is not the kind of race any legislation will have meaningful effect on.
The race is on and you better commit to a faction that may deliver.
> This race may even be the answer to the Fermi paradox
The mostly unchallenged popular notion that fleshy human intelligence will still be running the show 100s – let alone 1000s – of years from now is very naive. We're nearing the end of the human supremacy, though most of us won't live to see that end.
To be fair fleshy human intelligence has hardly been running the show any more than a bear eating a salmon out of a river thus far. We’d like to consider we can control the world yet any data scientist will tell you what we actually control and understand is very little, or at best a sweeping oversimplification of this complex world.
> has hardly been running the show any more than a bear eating a salmon out of a river thus far
Exactly. We're on an super-linear growth curve. Looking at the earth on a timescale since biological life emerged, it would seem that intelligent squishy life appeared for a split second and then an entirely artificial intelligent life took over from then on.
I agree with the analysis though LLMs might not be the specific tech that makes it.
I think the truly humanity-respecting solution is to figure out and implement a stable and fair-er 50% post-labor society before we actually get there. UBI (not the unmanaged flash layer) sounds like one part of the solution.
Trying to preserve employment is demonstrably wrong when already today there are too few unskilled jobs that make economic sense for the large amount of humans unable to learn more complex tasks.
> The race is on and you better commit to a faction that may deliver.
How does that help?
The giant does not care whether the ant he steps on worships him or not. Regardless of the autonomy or not of the AI, why should the controllers help you?
Due to the limitations of gradient descent and training data we are limited in the architectures that are viable. All the top LLM's are decoder-only for efficiency reasons and all models train on the production of text because we are not able to train on the thoughts behind the text.
Something that often gives me pause is the consideration that it is actually possible to come up with an architecture which has a good chance of being capable of being an AGI (RNNs, transformers etc as dynamical systems) but the model weights that would allow it to happen cannot be found because gradient descent will fail or not even be viable.
It is currently possible to use ZKP's to set up via a central authority a digital cash system where the bank notes are all anonymous and all transfers are anonymous.
The central authority in this scenario cannot discriminate between transactions - any function that would compare two or more transactions cannot glean any useful information that would allow to discriminate. And and security of the anonymity of past transactions will be reducible to the security of the cryptographic hash function used (the next best thing to Information-theoretic security). As for forging money, depending on what ZKP approach is used even a quantum computer will be insufficient.
The central authority can still print money and can obviously shut the entire system down.
It is interesting to ponder whether or not some government will decide to take such a step and surrender all control (except for the nuclear option) over how their currency is used. It will certainly boost demand for the currency.
Do you have any recommended references on this subject? Seems like this sort of system would be able to obfuscate a lot of metadata that can be used to deanonymize activity. Very interesting.
Tornado Cash does this. And you can find articles on how it functions online. You can even read the smart contract that directly implements it.
Roughly: you have 2 secrets that you hash together and the central authority adds the result you disclosed to a list (either to print money or as part of a transaction to transfer money). To spend a note you reveal the hash of one of the secrets (to be added to a list of nullifiers to prevent double spend) and you do ZKP to demonstrate that you possess both of the secrets to *some* note from the public hash list and that the nullifier for that note is what you claim it is. Central authority rejects if nullifier is present in the list.
There are some other approaches to such a system, I believe the Tornado Cash one is the most elegant though it limits you to a discrete number of note denominations.
Note that the proof system Tornado Cash uses is not secure to a quantum computer and such a device will allow to "print money" - in reality, drain the smart contract.
mRNA can be used as indirect vaccine, by manipulating cellular factories to produce a foreign object that draws the attention of the immune system.
There is no known pathway of persistence in this mechanism, and there is good reason to believe such a mechanism cannot exist as it would be a highly versatile attack vector and viruses would have evolved to be much more virulent by using it. Retroviruses do complex work to infiltrate the DNA repository of a cell, if all it took is to drop some mRNA it would have been exploited and 'patched' a billion years ago.
And if some of that mRNA happens to land in cells that don't readily get replaced, like heart muscle, what happens next?
>There is no known pathway of persistence in this mechanism
No known, and published in a paper, pathway. Engineering is about using things in practice we don't yet understand. They didn't know why suspension bridges kept falling down, so they just used a "safety factor" to decide how much to over-build a structure vs its theoretical capacity.
We have no idea what the long term consequences are of this gene treatment. It's not possible to have 10 year or more data on something just cooked up in the past 3 years.
