Bing's ChatGPT experiment is deeply flawed, and is the future of search

2»

Comments

  • Reply 21 of 29
    danoxdanox Posts: 3,887member
    tundraboy said:
    DAalseth said:
    Much potential, but yes there is a LONG way to go before you can trust the results. Part of the problem of course, is that people tend to ask ambiguous questions. What is the square root of 64 is fine. What is the best electric car, will likely produce unreliable results. I read a piece on this a while back where they started asking AIs questions with no correct answer, How to catch a bigfoot? Who invented nuclear fusion in the 1800s? That sort of thing. The results were pure trash, but they had citations and everything. There’s a lot of refinement that needs to be done before we can trust the results.
    When we humans engage in conversation, we are constantly trying to size up what the other person's thoughts are, especially their intent in talking with us.  Where are they coming from, so to speak.  What could possibly interest them and what would not.  And so on.  In short we are trying to peer into the other person's mind in the effort to make communication more effective.  Psychologists call that "Theory of Mind" and we engage our full cognitive toolkit, both logical and emotional, using visual and auditory queues, as well as situational and contextual awareness, to form an accurate theory of mind.

     AI doesn't have theory of mind.  It doesn't even have its own mind, which is the main prerequisite for forming a theory of another person's mind.  That AI is stupid, literally mindless, is no surprise.  The simulation of intelligence, no matter how authentic it looks, is not intelligence.

    The human mind, corvus, primates, orcas, octopus, etc. etc. do million upon millions of calculations per second most are done without even thinking about it, AI today is but a dream 20-50 a 100 years into the future?. The only thing I will agree on, is that when it happens, it will accelerate so fast that the humans and everything else may be extinct soon after. (And the extinction of many species will happen long before that, due to man).
    watto_cobraAlex1N
     2Likes 0Dislikes 0Informatives
  • Reply 22 of 29
    Is the upcoming Apple event going to provide any response from Apple to GPT?

    https://appleinsider.com/articles/23/02/07/apple-to-hold-first-in-person-ai-summit-in-years
    watto_cobra
     1Like 0Dislikes 0Informatives
  • Reply 23 of 29
    dewme said:
    I’ve been waiting forever for this to happen. I loathe search engines as they are because they are keyword based. This is not how humans seek information.

    Sure, this AI has a way to go but it will get better quickly now that M$ and to a lesser extent Google are onboard.

    Google however will not be going full boar into this development though. They’ll give the impression that they are but ultimately it’s too much of a risk to their business model for them to take it seriously. M$ and Apple have a vested interest in this route. The difference is that Apple is doing it on device not on browser.

    Those complaining about Siri aren’t seeing the forest for the trees. Apple is working behind the scenes to get the technology there before it’s fully unleashed on the world. They’re giving us snippets of what Siri is capable of. I’d say in 5 years Siri will outperform all other AI systems because the backends will be far more powerful.
    I’m not seeing where your faith in Apple’s ability to overcome obstacles that companies having far greater knowledge and experience in information science, library science, classification systems, and understanding temporal and contextual associative relationships comes from. Google and Amazon have far more experience and expertise in these areas, many of which have been evolving since the 17th century. The scientific and mathematical basis for some of the association and classification algorithms were derived well before computers were available to actually execute the logic in real time and at scale. It’s similar to many of the advancements in computer architecture which were designed in the 1950s and 1960s but only became relevant and usable when the implementations could be done at VLSI levels at the micro and nano level.

    In simple terms, coming up with solutions using AI, ML, big data, etc., aren’t a matter of cleverly pulling a rabbit out of a hat. The “hat” that hosts the rabbit is very deep in sciences that have existed for quite some time, hundreds of years in some case, and requires specific domain  experience and knowledge that has not been the primary focus of device makers, unless the device is merely a portal into the services that are generating the revenue. 

    Apple has been a product company that sells devices. They are relatively new to the services businesses that lean on the technologies contributing to AI and those that require access to big data. Apple has purposely limited its ability to delve too deeply into associating privileged data that it receives from its customer base, which puts them behind the curve in many ways compared to say Google or Amazon. I don’t think we need to come up with rationale to explain Siri’s slow development, much less portray it as if Apple is “holding back” to make sure they have it “perfected” before showing off its “true magic” to the world. Apple is still learning, plain and simple. They didn’t invent the technology, they bought it. That’s perfectly fine. But now that it’s theirs they’ve been doing their homework to understand the science behind it better and see what kind of Apple-specific secret sauce they can add to it without stepping on their customer’s toes, like some other companies have no problem doing.

    The whole notion of Apple having any overwhelming competitive advantage on “back end” technology, much less implementation, is wildly optimistic. They aren’t close to being in that position now, but who knows, they may get there eventually. They have a massive number of very smart people and nearly unlimited resources. 
    My faith in Apple comes from their past releases.

    Photos is doing more in terms of OCR than has ever been done before. If search providers like Google, M$, or even Amazon had the skills why have they only just started releasing products?

    Apple has access to all sorts of information, much of it the same places as M$ and Google. Most importantly though, Apple has the drive to further this technology.

    Apple has moved us to accept AR which gives it access to data that Google doesn’t/can’t leverage. Hell, compared to Google Maps, Apple Maps is in my experience a far more accurate system. I know many people disagree but I can tell you that Apple Maps gives me more streets in my city than Google Maps does. If Google has more skills than Apple why is t it delivering?
    Alex1N
     1Like 0Dislikes 0Informatives
  • Reply 25 of 29
    So many questions regarding integrity and worldview here.

    a computer program that simulates intelligence compiling information from…?

    and who decides what’s right or wrong in topics of political/divisive nature?

    Everyone remembers big tech shutting down everyone who did not toe the political lines concerning various aspects of COVID and associated government responses. And now it turns out the dissenters were proven correct. 

    Hopefully this doesn’t become a political power tool, you know, once it’s actually useful. 

    muthuk_vanalingambyronl
     2Likes 0Dislikes 0Informatives
  • Reply 26 of 29
    Just like Siri….unusable
     0Likes 0Dislikes 0Informatives
  • Reply 27 of 29
    globby said:
    By far the scariest thing is ChatG’s ability and demonstrated tendency to make up nonexistent sources and studies to back up her answers. Not what we need for search results. 
    Welcome in The Matrix
     0Likes 0Dislikes 0Informatives
  • Reply 28 of 29
    dewmedewme Posts: 6,098member
    dewme said:
    I’ve been waiting forever for this to happen. I loathe search engines as they are because they are keyword based. This is not how humans seek information.

    Sure, this AI has a way to go but it will get better quickly now that M$ and to a lesser extent Google are onboard.

    Google however will not be going full boar into this development though. They’ll give the impression that they are but ultimately it’s too much of a risk to their business model for them to take it seriously. M$ and Apple have a vested interest in this route. The difference is that Apple is doing it on device not on browser.

    Those complaining about Siri aren’t seeing the forest for the trees. Apple is working behind the scenes to get the technology there before it’s fully unleashed on the world. They’re giving us snippets of what Siri is capable of. I’d say in 5 years Siri will outperform all other AI systems because the backends will be far more powerful.
    I’m not seeing where your faith in Apple’s ability to overcome obstacles that companies having far greater knowledge and experience in information science, library science, classification systems, and understanding temporal and contextual associative relationships comes from. Google and Amazon have far more experience and expertise in these areas, many of which have been evolving since the 17th century. The scientific and mathematical basis for some of the association and classification algorithms were derived well before computers were available to actually execute the logic in real time and at scale. It’s similar to many of the advancements in computer architecture which were designed in the 1950s and 1960s but only became relevant and usable when the implementations could be done at VLSI levels at the micro and nano level.

    In simple terms, coming up with solutions using AI, ML, big data, etc., aren’t a matter of cleverly pulling a rabbit out of a hat. The “hat” that hosts the rabbit is very deep in sciences that have existed for quite some time, hundreds of years in some case, and requires specific domain  experience and knowledge that has not been the primary focus of device makers, unless the device is merely a portal into the services that are generating the revenue. 

    Apple has been a product company that sells devices. They are relatively new to the services businesses that lean on the technologies contributing to AI and those that require access to big data. Apple has purposely limited its ability to delve too deeply into associating privileged data that it receives from its customer base, which puts them behind the curve in many ways compared to say Google or Amazon. I don’t think we need to come up with rationale to explain Siri’s slow development, much less portray it as if Apple is “holding back” to make sure they have it “perfected” before showing off its “true magic” to the world. Apple is still learning, plain and simple. They didn’t invent the technology, they bought it. That’s perfectly fine. But now that it’s theirs they’ve been doing their homework to understand the science behind it better and see what kind of Apple-specific secret sauce they can add to it without stepping on their customer’s toes, like some other companies have no problem doing.

    The whole notion of Apple having any overwhelming competitive advantage on “back end” technology, much less implementation, is wildly optimistic. They aren’t close to being in that position now, but who knows, they may get there eventually. They have a massive number of very smart people and nearly unlimited resources. 
    My faith in Apple comes from their past releases.

    Photos is doing more in terms of OCR than has ever been done before. If search providers like Google, M$, or even Amazon had the skills why have they only just started releasing products?

    Apple has access to all sorts of information, much of it the same places as M$ and Google. Most importantly though, Apple has the drive to further this technology.

    Apple has moved us to accept AR which gives it access to data that Google doesn’t/can’t leverage. Hell, compared to Google Maps, Apple Maps is in my experience a far more accurate system. I know many people disagree but I can tell you that Apple Maps gives me more streets in my city than Google Maps does. If Google has more skills than Apple why is t it delivering?
    Apple, Google, Amazon, Microsoft, Meta, etc., are not in a battle to achieve product-level and service-level parity in every market that each one competes. Each company has a specialization that distinguishes them from all the others. Apple’s distinguishing capability is in product design and product delivery at scale, which they do extremely well at and are richly rewarded for doing so. Google’s primary specialization is their ad business which manifests itself through their search, mapping, media (YouTube), analytics, cataloging, etc.. Microsoft’s specialization is in software and cloud services. Amazon’s specialization is in online sales, web services, etc.

    All of these companies also do many other things, like building hardware and software products, primary research, infrastructure and architecture, media streaming, running data and cloud service centers, building middleware for sale, etc., but mostly to prop up their primary focus and areas of specialization. As each of these mega companies move toward the application of AI it’s safe to assume that they will be trying to leverage AI in ways that complement their primary business focus. 

    I think it comes down to answering the question: “How will Apple/Google/Microsoft/et al, best leverage AI in the products they sell?” This isn’t a race to see who can deliver the “Best AI” that delivers on all of the promises of AI to the world, with a winner take all reward handed out. It’ll probably play out much like Siri vs the other voice based assistants/query tools. The implementation of AI in Apple products will be scoped to things that Apple deems to be essential, no more and no less. We’ve already seen what happens when someone over-delivers on technology like this, I.e., with Amazon’s Alexa. It ends up with a lot of superfluous fluff like Fart Skills and the company building it loses a lot of money on the feature. But from a direct tech-vs-tech comparison, the objectively measurable “better” implementation is the one that incorporates the advantages owned by its builder. But that may not matter. Building the right thing matters just as much as building the thing right.

    I don’t believe Apple has a monopoly on innovation, motivation, intelligent employees, intellectual property, technology, or market access versus any of the companies trying to bring AI into products. In my opinion Apple’s biggest advantages since the beginning of the SJ 2.0 era have been 1) Building the Right Things, 2) Execution, and 3) Delivery at Scale. I’m sure there are many more and there have been a couple of minor stumbles and not everyone agrees with the Right Things all of the time (e.g., trash can MacPro), but the bottom line success speaks for itself. Steve Jobs was obviously stronger in the Right Things department, but nobody comes close to executing as well as Apple has with Tim Cook at the helm. We saw what happened in the last quarter when Delivery at Scale suffered a hiccup. Apple’s strengths are still fully intact, at least for the time being. We’ll have to wait and see what happens with some of the rumored things like Apple Car and AR/AI that don’t intuitively play into Apple’s strengths based on their current highly successful portfolio.
    edited February 2023
    muthuk_vanalingamAlex1N
     2Likes 0Dislikes 0Informatives
Sign In or Register to comment.