I think what your getting at is basically the idea that LLMs will never be "intelligent" in any meaningful sense of the word. They're extremely effective token prediction algorithms, and they seem to be confirming that intelligence isn't dependent solely on predicting the next token.
Lacking measured responses is much the same as lacking consistent principles or defining ones own goals. Those are all fundamentally different than predicting what comes next in a few thousand or even a million token long chain of context.
Indeed. One could argue that the LLMs will keep on improving and they would be correct. But they would not improve in ways that make them a good independent agent safe for real world. Richard Sutton got a lot of disagreeing comments when he said on Dwarkesh Patel podcast that LLMs are not bitter-lesson (https://en.wikipedia.org/wiki/Bitter_lesson) pilled. I believe he is right. His argument being, any technique that relies on human generated data is bound to have limitations and issues that get harder and harder to maintain/scale over time (as opposed to bitter lesson pilled approaches that learn truly first hand from feedback)
I disagree with Sutton that a main issue is using human generated data. We humans are trained on that and we don't run into such issues.
I expect the problem is more structural to how the LLMs, and other ML approaches, actually work. Being disembodied algorithms trying to break all knowledge down to a complex web of probabilities, and assuming that anything predicting based only on those quantified data, seems hugely limiting and at odds with how human intelligence seems to work.
Sutton actually argues that we do not train on data, we train on experiences. We try things and see what works when/where and formulate views based on that. But I agree with your later point about training such a way is hugely limiting, a limit not faced by humans
Based on the comments here, it sounds like repurposing a boolean "verbose" mode and having that verbose mode actually be multi-state is confusing.
It might be worth considering a "verbose level" type setting with a selection of levels that describe the level of verbosity. Effectively, use a select menu instead of a boolean when one boolean state is actually multiple nested states.
Edit: I realised my use of "verbose" and "verbosity" here is it self ironically verbose, sorry!
Killing the idea of basic income would be a good thing. It will never work and would leave a society in much the same situation as other past attempts at Marxism.
...and when your labour, and the labour of 90+% of all humans on the planet have no economic value, we'll do what?
Continue to avoid exploring obvious solutions because certain words have been made into epithets, or failed previously because they were solving future (now imminent) problems?
If we don't fundamentally change our economic system its simple, we're all screwed.
If we have a system depending on trading our labor for money to pay for stuff, and the value of our labor goes to zero, we need a different system.
We can't paper over that fundamental crack by giving governments even more power to decide what every person "needs" and send out resources accordingly.
There are so many problems in that system. How do we actually decide prices when every consumer has the same base level of money to spend? How does the government decide what we all need or deserve? How do we avoid the corruption taking over that massive power granted to dole out resources? Are we just living in a feudal state again? Does the government need to control the means of production to keep such a system stable?
Price inflation isn't the only type of inflation. Monetary inflation was historically the "inflation" people focused on, and given that Ireland runs a deficit in their state budget this would be adding to the debt and inflating the money supply.
What you're describing then would only be a short term chicken and egg problem.
If, once established, a thriving art scene generates value by attracting tourism, wealthy individuals who want to patronize the arts, etc then the artists would be able to charge enough to fund themselves and potentially do very well for themselves.
In that scenario we'd only need to fully subsidize artists for a short period of time, the subsidy law should have an expiry date.
Its mincing words a bit, but an attack targeting drug cartel assets wouldn't necessarily be viewed as a war with Mexico. It could lead to that for sure, and the Mexican government could declare it an act of war, but we did just see the US literally invade a foreign country and arrest their sitting leader without war being declared on either side.
We declared war on drugs and on terror, maybe AIDs and Covid as well? Though you're right, we haven't declared war on another state since WWII despite being in multiple wars over that time.
I assumed when you wrote "war being declared" you meant in Constitutional sense which reserves to Congress the power to declare war.
Not in the metaphorical "war on poverty" sort of way.
FWIW, examples in addition to Maduro are Aguinaldo (Philippines), Noriega (Panama), Hussein (Iraq), and Aristide (Haiti).
(Technically speaking, the US didn't recognize Philippine independence so didn't consider Aguinaldo to be its president, but instead a rightful cession from the Kingdom of Spain due to the Treaty of Paris, which ended the Spanish–American War, where the US had made a formal declaration of war.)
(Also, the US says Aristide's departure was voluntary.)
You’re missing the point. Absolutely cartel violence impacts all types of people in the US and Mexico but large scale brutal violence that is usually saved for Mexico since unfortunately the Mexican federal government does not have control in most of the regions.
There is a huge difference between a one off gang killing in the US and someone taking a whole grey hound bus and burying the bodies in the desert.
The world obviously doesn't stop st the US border. The point in this thread was that the attacks on buses full of people have, so far stopped at the US border and that it would be a huge, and dangerous, escalation should that change.
git didn't succeed based on the mental model. It got a foot in the door with better tooling and developer experience then blew the door open when GitHub found a way to productize it.
Lacking measured responses is much the same as lacking consistent principles or defining ones own goals. Those are all fundamentally different than predicting what comes next in a few thousand or even a million token long chain of context.
reply