nopinsight

22 ms·

271.
by nopinsight 3y ago
Current LLMs clearly have much coarser models of the world than humans do. Their training data do not include as many modalities and training data with spatiotemporal dimensions are inadequate. Up and coming multimodal models are changing t
272.
by nopinsight 3y ago
GPT-4 should do significantly better, but still worse than expert humans, on the tasks you mentioned. They would also benefit from good prompting, such as those in - https://lilianweng.github.io/posts/2023-03-15-prompt-
273.
by nopinsight 3y ago
Current LLMs likely have multiple world models with varying qualities. Good prompts are required to activate suitable models for each task. 1. Did you use GPT-3.5 (free version) or GPT-4 (paid one) to form the judgment? Their performances o
274.
by nopinsight 3y ago
> Nothing of real value imho. Without the right target market, business model, and effective methods to reach customers, the most brilliant pair of shoes or piece of code can be useless (unless someone works to repurpose them as art or a
275.
by nopinsight 3y ago
> You still need to know what you’re doing though, you can’t blindly copy and paste. Agreed. > Like a purse, millions would pay money for it, some even pay thousands. Expensive purses have intangible value for some. They are often bou
276.
by nopinsight 3y ago
More details in this blog post by a Wharton professor: https://www.oneusefulthing.org/p/centaurs-and-cyborgs-on-the... My questions to naysayers: * Do you or anyone you know use GPT-4 (not the free GPT-3.5) to do produ
277.
by nopinsight 3y ago
Accidents and health problems are not completely predictable for an individual, but the rates are for different groups of people. As long as it’s within their power to choose, they should accept the consequences that come with their choices
278.
by nopinsight 3y ago
The data might not be available at the time. And the person did not have a choice when they were a child. Is the above question rhetorical or do you lean more libertarian?
279.
by nopinsight 3y ago
I agree than tracking detailed purchase data can lead to serious discrimination and inequity especially in a dictatorship. The use case in your example, however, seems fair: * The person has a choice whether to consume fast food regularly.
280.
by nopinsight 3y ago
I was addressing your statements regarding the capabilities of neural AI: > Anyway, all known AI techniques struggle with common sense reasoning. The majority of ambiguities in natural language are distinct and uncorrelated. Combining kn
281.
by nopinsight 3y ago
LLMs still do not match humans on Winograd Schema Challenge but the top ones already get pretty close. The current top model, Vega v2, is already at 98.6 (compared to human’s 100.0). https://super.gluebenchmark.com/leaderbo
282.
by nopinsight 3y ago
Articles from The Economist, The Wall Street Journal and The New York Times are better and also more expensive to produce than those from Buzzfeed, which justify their subscription fees. What if editors of such caliber could leverage AI-ass
283.
by nopinsight 3y ago
>> Unfortunately, I suspect we will get a two-tiered system, where the "middle class" (whether that's disappearing is another question) can afford human content/human support/etc, and the working class are for
284.
by nopinsight 3y ago
Geoffrey Hinton, Andrew Ng, and quite a few other top AI researchers believe that current LLMs (and incoming waves of multimodal LFMs) learn world models; they are not simply 'stochastic parrots'. If one feeds GPT-4 a novel proble
285.
If it can be designed on a computer, it can be built by robots(economist.com)
112 points by nopinsight 3y ago|99 comments
286.
by nopinsight 3y ago
I recall my former professor at UT Austin, J Strother Moore, sharing how Intel utilized an Automatic Theorem Prover he helped create to spot bugs in processors. But when engineers tweaked the design just a little, thinking it was a no-brain
287.
by nopinsight 3y ago
I think humans need to move up the levels of abstraction and master specific domain knowledge. Barring AGI, there will still be need for humans to elicit and learn the requirements from customers and other stakeholders, then translate those
288.
by nopinsight 3y ago
GPT-3.5, GPT-4, and similar are fairly new. There are many uses of the technology that remain unexplored by product people, not to mention the technology is getting more advanced and has more capabilities by the week (or day or month, depen
289.
by nopinsight 3y ago
We do have regulations and laws to control genetic modifications of pathogens. They are done in highly secure labs and the access is not widely available to anyone. If a pathogen more deadly than Covid starts to spread, eg like Ebola or Sma
290.
by nopinsight 3y ago
Genetic modifications could potentially cause havoc in the long run as well, but it's much more likely we have time to detect and thwart their threats. The major difference is speed . Even if we knew how to create a new species of sup
291.
by nopinsight 3y ago
LLM is already a misnomer. Many of the latest models are better called LFMs (Large Foundation Models). They have multimodal capabilities. Some can even handle sensory input humans can't. Another comment already links to demos and paper
292.
by nopinsight 3y ago
One does not need to be have super-human capabilities to foresee potential risks of the technology one is an expert in. Yes, the current version of AI is not capable of large scale harm by itself yet, but the plausible trajectory is worth w
293.
by nopinsight 3y ago
I'd say stronger evidence than the Othello paper is the ability to answer what-if questions coherently and plausibly. I just asked GPT-4 "What would happen to Northern Thailand if elephants behave like kangaroos?". The answer
294.
by nopinsight 3y ago
You are attributing naivety on a technical matter to many people who have done world class research and some of them are even a pioneer of their subfield of study. To do that level of research, one needs strong independent thinking skills a
295.
by nopinsight 3y ago
I’m not sure there is a way for someone to engineer a deadly virus while completely innoculating themselves from it. Short-term AI risk likely comes from a mix of malicious intent and further autonomy that causes harm the perpetrators did n
296.
by nopinsight 3y ago
Here’s why AI risks are real, even if our most advanced AI is merely a ‘language’ model: Language can represent thoughts and some world models. There is strong evidence that LLMs contain some representation of world models it learned from t
297.
by nopinsight 3y ago
LLM is already a misnomer. Latest versions are multimodal. Current versions can be used to build agents with limited autonomy. Future versions of LLMs are most likely capable of more independence. Even dumb viruses have caused catastrophic
298.
by nopinsight 3y ago
In interviews, Geoffrey Hinton and Yoshua Bengio certainly expressed serious concerns and even some plausible regret to their life’s work. They did not say anything that can be interpreted as your last sentence suggests at all.
299.
by nopinsight 3y ago
Even if it’s just Yoshua Bengio, Geoffrey Hinton, and Stuart Russell, we’d probably agree the risks are not negligible. There are quite a few researchers from Stanford, UC Berkeley, MIT, Carnegie Mellon, Oxford, Cambirdge, Imperial College,
300.
by nopinsight 3y ago
It’s worth noting also that many academics who signed the statement may face adverse issues like reputational risk as well as funding cut to their research programs if AI safety becomes an official policy. For a large number of them, these

More ›