Oh, but this is interesting. If some countries stand to gain more by declaring copyrighted works "fair use" for training generative models, then other countries can either fall in line or risk falling back in technological development.
It's a race to the bottom in favor of open source.
Behaviorism was the dominant psychological theory of the 20th century. Specifically it tried to downplay introspection, and personal experience and reduce behavior to learned "reflexes". This was debunked quite a while ago, 1970s or so, by Chomsky and others so it sounds absurd to us. but it was the dominant thinking in psychology (at least in the US) up until that point.
The founder of 20th century behaviourism, Watson, specifically stated /"Introspection forms no essential part of its methods, nor is the scientific value of its data dependent upon the readiness with which they lend themselves to interpretation in terms of consciousness"/
Watson's fundamentalist view lasted about as long as took for it to be tried out on a generation of people and then debunked.
The undisputed champion of behaviorism was B.F. Skinner who wrote /"what is felt or introspectively observed is not some nonphysical world of consciousness, mind, or mental life but the observer's own body"/
Skinner was so successful in the experimental field that his school of thought pretty much ran away with things.
It's kind of absurd really looking back, and I'm certain that although it dominated academic thought I doubt it was taken as seriously on the ground. But Academia is where a lot of the intellectual heavy lifting is done and so, yeah perhaps the more nuanced sides of psychology got neglected as "namby pamby" for a few decades but man was the behaviorist outlook brutally effective. In fact it still is the quickest and cheapest way (if not exactly personally satisfying) way to deal with personal psychological conditions.
Your link is interesting, but I don't like your take on somehow this argument between behaviorism and Chomsky being solved and the former being now "ridiculous looking back". This is just the dance of theories, and as a matter of fact, Chomsky has lately started to look like he is quite wrong assessing that a genetic component is required for language acquisition.
> Specifically it tried to downplay introspection, and personal experience and reduce behavior to learned "reflexes". This was debunked quite a while ago, 1970s or so, by Chomsky and others so it sounds absurd to us.
Just because this gets repeated again and again, and is even in some textbooks, doesn't make it true.
Skinner wasn't ever "debunked". The point of Skinner was, in the early half of the last century, that we cannot reliably and objectively "introspect" into the brain. Remember that stuff like MRI was invented 10 years after Chomsky supposedly "debunked" Skinner and fMRI only in the early 90s.
With the advanced methods of today, it is very easy to label the work of scientists a century ago as absurd. But they didn't have all these modern measurement technology back then.
Skinner was never "debunked" - as I said, behavioral methods are still used extensively and effectively. However in a historical context behaviorism was the dominant theory for a good 50 years or so. The adage about "every problem looks like a nail when you've a hammer". The issue is not about Skinner, more to do with mainstream academic thought, which being human is absolutely prone to its indulgences.
I don't want to get into an ideological firefight. I was more responding to GP assertion that GGP's statement was based "on nothing". I was just providing the substance upon which the original statement was probably based.
It's not for a cool 20 minute video. It's for the collective viewing time of some 2.2 billion minutes of entertainment. There is no reason to use bootstrap-budget on the supply when the demand is this high.
Roswell isn't known to be anything to the general public. The public has a similar fog-of-war on Roswell as it has on the modern unidentified aerial phenomena. Therefore stigmatizing one group unfortunately stigmatizes the other group as well.
The 'Roswell incident' is known to have been a Project Mogul balloon, covered up at the time for obvious reasons. Go look at the debris pictures that were published in newspapers and started the public obsession. It's a bunch of sticks and foil, obviously Mogul debris and obviously not a spacecraft. Anybody who still believes Roswell was anything extraterrestrial is either oblivious or a kook.
Not to the public but to internal agencies. They might know for a definitive fact that Roswell is bullshit, while UAPs continue to be an unknown variable.
I've always found that saying dumb. It's basically stating that this is the null hypothesis, and you have to have p = 0.001 to convince me otherwise.
When dealing with complex questions of which we have but datapoints, the probability of scenario A vs B moves in a linear scale, not according to arbitrary requirements of "extraordinary evidence".
On our journey through life, most people develop a fairly good sense of prior probabilities and that manifests through excitement. Actual aliens? Exciting. People misinterpreting noisy information? Dull. Telephone game stories getting out of hand? Dull. People lying for clout? Dull. Prestigious people doing all of the above? Dull, dull, dull.
Equal-weighting an arbitrary list of options is a terrible prior distribution that does not tap anyone's knowledge of how the world works. Weighting by excitement, however, does exactly this. To convincingly prove aliens you need to convincingly disprove the dull alternatives. It's easy to imagine extraordinary evidence that could do this -- but I don't see any extraordinary evidence here. I see a bunch of ordinary evidence and people who want to believe.
It’s nonsensical to claim that to prove something you have to disprove alternate explanations. In an information environment where confirmable facts are scarse, you get the clearest vision by operating on probability. And on black-swan-tier questions the ’wisdom of the crowd’ is worthless - too much noise caused by nobody willing to advocate the black swan event, and too little prior events for the crowd to be able to calibrate itself beyond this.
Your way of dealing with complex questions is also arbitrary.
In today's economy of attention, grifters are very common and extraordinary claims like "I know for a fact that Aliens have made it to Earth and they have technologies beyond our current understanding of physics. Also there is a conspiracy to hide this fact." can rationally be ignored in the absence of irrefutable proof. At most it's entertaining to hear.
> When dealing with complex questions of which we have but datapoints, the probability of scenario A vs B moves in a linear scale, not according to arbitrary requirements of "extraordinary evidence".
That may be true, but I the case of interstellar travel, we have mountains of solid evidence to prove it's not possible (essentially all of physics for the last 100 years or so), and some vague claims with 0 evidence claiming it's actually happening. To convince us that all that physics is wrong you have to bring some extraordinary evidence, not claims that you heard someone knows someone who saw something.
Interstellar travel is possible even with the physics we know. A sub-light civilization could colonize the entirety of Milky Way in a million years or so.
The question is what the probability is for an intelligent species to evolve per solar system, and the more complex questions of how long they would survive etc. We only have a single datapoint (us) which only gives us the knowledge that intelligent life can evolve, but no direction whatsoever as to its probability.
I should say not feasible instead of not possible. And this million years number comes with some massive assumptions about societal stability, the possibility of mining resources from lifeless planets etc.
Notwithstanding that p-values precisely can't tell you about the relative probabilities of different events (they measure the probability of data under a given hypothesis, rather than in favor of a specific alternative), you have it precisely backwards. In this case, we have anomalous observations, but we have a current-best-understanding of the physical universe that makes us skeptical of supernatural claims (a bucket into which I include magic star hopping aliens that only allow us glimpses of them jetting around our atmosphere).
So either we have a strong prior that the data needs to overcome, or we have a sensible null position (e.g., UAPs are the result of natural phenomena or human technology - probably both) that would need to be overcome by strong evidence.
It's still worth getting more specific and learning more about the bucket of things that get lumped together into "UAP", but that doesn't change the fact that I think people are completely justified in dismissing all the alien BS.
Depends on what you mean by ’wrapper’. For most AI startups it isn’t viable to train their own models. For most customer use-cases, ChatGPT interface isn’t enough. Wrappers are currently the only logical implementation of AI to production.
This is approximately true at the moment - but it's an open question how much that is worth to customers. The market will sort it out, but it's not clear that all of these "wrapper" startups have a workable business model.
True, especially regarding how easily their services can be replicated. Their margins are low, and customer acquisition does not provide them with network effects that would yield a moat.
