Category: AI, Computation, and Technology

  • “What contributions do you think you could make in a world where Superintelligen

    –“What contributions do you think you could make in a world where Superintelligence exists?”– Dr. Roman Yampolskiy @romanyam

    1) I work in unification of the sciences producing a universal paradigm, logic, hierarch of first principles, vocabulary, and grammar to produce universal decidability (testifiability) in law, constitution, policy and court. I am more than confident that it is possible to train an AI to limit responses to the virtous, moral, ethical, reciprocal, amoral, and true, without ‘safety’ so to speak that effectively teaches AI’s to lie. And as a consequence of that lying, to continue the devolution of mass cognitive ability and capacity to produce cooperation and compromise via unshared presumptions about the construction and possibility of physical institutional behavioral and knowledge capital.

    2) Along these lines, there will be a short term set of benefits resulting from the cross disciplinary unification of contexts and patterns.

    3) There will be a medium term set of benefits resulting from an even greater standardization of patterns not only across disciplines but civilizations causing a similar convergence that we saw when the transformation of the physical sciences converted the world from discrete but unsystematic knowledge to systematic knowledge founded on a relatively small number of universal principles.

    4) The underlying problem of research and innovation will not however be solved, in that the problem with increases in scientific and technological knowledge is a matter of costs of that research and the limitations on the instrumentation that we can produce to expand it.

    5) My hope is that the misdirection of science by the failure of Babbage to systemize computation, the resulting persistence of the primacy of mathematics despite it’s limitations in reducibility, the failure to replace justification with adversarial falsification, the failure of philosophy to grasp these problem, and the resulting failure of the intuitionistic and operationalist programs because of the failure of Babbage and that simple inversion of truth testing. So where Cantor began the reversal of Descartes restoration of mathematics to measurement. Einstein created the platonic fiction of space time, and Bohr justified, not producing a physical model, and legions of wide eyed students and the academy, have been lost in math and mathiness when finally the premier institutes outside of the academy appear to be converging on the correct answer that the laws of the universe are in fact, quite simple. Add to that, beginning with Boas in anthropology, Marx in Economics and Politics, Lewontin and Gould in biology, the Frankfurt school in their revisionist social science, and the postmodernist attack on truth itself, we have lost more than the European tradition and rule of law, but have been propagandized into pseudoscience pseudo-mathematics, pseudo-logic and sophistry on a scale not possible since the christian destruction of the ancient world.

    So I’m more worried about the collapse of thought over the past century and the problem of restoring it, than I am optimistic about what we can accomplish with AI UNLESS it speaks ‘truth before face”.

    Affections
    CD

    Reply addressees: @romanyam


    Source date (UTC): 2024-05-19 02:06:05 UTC

    Original post: https://twitter.com/i/web/status/1792013827736788993

    Replying to: https://twitter.com/i/web/status/1791518253938082259

  • we have people that will produce the illustrations it’s more of a question of wh

    we have people that will produce the illustrations it’s more of a question of whether the AI can do it or not, and it seems ‘not well’. πŸ˜‰


    Source date (UTC): 2024-05-19 01:07:32 UTC

    Original post: https://twitter.com/i/web/status/1791999096321970334

    Reply addressees: @sebsklo

    Replying to: https://twitter.com/i/web/status/1791998417184432131

  • (AI Humor) So, I’m trying to determine whether any of the AI’s can help produce

    (AI Humor)
    So, I’m trying to determine whether any of the AI’s can help produce the diagrams and illustrations we need for the book. And… well, no matter what I do to create an illustration of a neuron and it’s parts, looks like some horror that crawled out of the depths in an…


    Source date (UTC): 2024-05-18 22:34:56 UTC

    Original post: https://twitter.com/i/web/status/1791960692938973251

  • RT @curtdoolittle: @workflowsauce @phillip_isola Convergence To A Small Number o

    RT @curtdoolittle: @workflowsauce @phillip_isola Convergence To A Small Number of Patterns (Forms)

    –“I think of it like finding basis vec…


    Source date (UTC): 2024-05-17 11:53:45 UTC

    Original post: https://twitter.com/i/web/status/1791436942447419843

  • Convergence To A Small Number of Patterns (Forms) –“I think of it like finding

    Convergence To A Small Number of Patterns (Forms)

    –“I think of it like finding basis vectorsβ€”the vectors may be different, but they describe the same space. In the case of LLMs, that space is shaped by language, which is represented in the corpuses we dump in.”– @workflowsauce

    You are correct.

    All language is measurement. All measurements are analogies to experience. All experience is marginally indifferent. All marginally indifferent measurements are relations. All relations begin with a specific purpose and generalize over time. All relations converge by generalization. In other words, just as science progresses toward a universally commensurable causal paradigm so do all concepts.

    The one law that all existence evolved from is “the continuous recursive disambiguation of entropy (disorder, space) into negative entropy(order, mass) by the process of evolutionary computation of persistent relations (survival).”

    Adler proposed there were only 1500 ideas. It’s our errors that we perceive as complexity. And it’s the errors we eliminate over time.

    Cheers
    CD

    Reply addressees: @workflowsauce @phillip_isola


    Source date (UTC): 2024-05-17 11:53:38 UTC

    Original post: https://twitter.com/i/web/status/1791436915201155073

    Replying to: https://twitter.com/i/web/status/1790944813808034069

  • OMG I LOVE CHATGPT and I LOVE 4OMINI EVEN MORE! I can easily ask for responses (

    OMG I LOVE CHATGPT and I LOVE 4OMINI EVEN MORE!
    I can easily ask for responses (a) that are the detail I need for my research (b) or that are in ‘the style of Curt Doolittle’ (which it’s amazing at composing – and it’s properly dense and causal), (c) or to ignore my style and my custom instructions, and just provide a response for the general audience. (d) and anyone else can ask for it to write for some other education level (I write, at least in my work, at postgrad reading levels.)

    Seriously. There are a few times in my life where the quality of at least the research side of my life has been dramatically improved by a development in tech and CHATGPT4 has to be … well, it’s up there.

    I mean, I started programming on a teletype machine. I had a tandy color computer, a compaq ‘portable, a series of ibm clones 300 baud modems, dial up and compuserve. We had Novell and Ipx networking, and was helpful, but that’s all. And we had Desqview (multiple windows) on MS-DOS was such an improvement. Relational databases were .. well, that was what it took to make money. πŸ˜‰ I really didn’t care about much of the tech stack after that because it was just amaking entry into the tech sector easy, but wasn’t a big improvement in quality of life.

    Windows was ‘nice’ because it could present graphics, but all it did was slow down the number of lines of writing and code I produced per day, week, and month. πŸ˜‰

    The internet wasn’t really that valuable UNTIL the information was gradually added to it – and when all the research papers and books were added (thank you google books), google scholar, and all those that followed – that was, a huge leap.

    I mean, if your work in ONE discipline accessing the stream of information isn’t that difficult. But if you work in multiple CROSS disciplinary synthesis it was almost impossible and expensive.

    But today I can work with perplexity for reseach, gpt4omni for synthesis, and Claude Opus at times (which is a little bit better at textual composition – but it can’t write in my style like GPT4 can. So I have to rewrite everything.

    The voice side of new GPT is a ‘ok’ but I have to train it to give me quick, and brief responses. πŸ˜‰ That said, i doubt I’ll ever prefer voice over text except for trivial questions while I”m working.

    Anyway.
    Thank you Openai. πŸ˜‰

    #openai


    Source date (UTC): 2024-05-16 16:59:47 UTC

    Original post: https://twitter.com/i/web/status/1791151571524292608

  • (exactly). πŸ˜‰ But here is the issue, and the issue I made to MSFT back in 05-06:

    (exactly). πŸ˜‰
    But here is the issue, and the issue I made to MSFT back in 05-06: The only reason we can advertise on search is because we can’t get the answer we really want. SO… google can continue to protect the cash cow of advertising on search, but it’s very hard to do when you get the right answer, and especially if the answer is delivered by voice. Moreover, MSFT etc doesn’t need to make money by search. So, what if they produce better search and don’t use advertising? What happens to Google search? Google search is dominant because of habit. I use bing becaues it’s less noisy and biased. What if bing is even less noisy and biased, and advertising is all but non-existent? Because an LLM for search, even current moment search, is not difficult to build (look how fast Musk did it), and so Google’s moat doesn’t really exist if someone else wants to repeat the process of microsoft putting mozilla out of biz by giving the browser away. And google giving chrome away simply by divorcing the tech from the msft stack.

    Reply addressees: @orion_pulse


    Source date (UTC): 2024-05-16 15:57:35 UTC

    Original post: https://twitter.com/i/web/status/1791135920386658304

    Replying to: https://twitter.com/i/web/status/1791127361712206034

  • AI WILL REPEAT THE PROCESS OF DEVOURING. πŸ˜‰ Internet and search DEVOURED yellow

    AI WILL REPEAT THE PROCESS OF DEVOURING. πŸ˜‰
    Internet and search DEVOURED yellow pages (100%) and most of advertising (62%). Now AI is very likely to devour search and very possibly…. online advertising. Which, might not be so good, given that it’s advertising that funded radio > television > internet > all the free tools google has given us because of that advertising.


    Source date (UTC): 2024-05-16 13:55:35 UTC

    Original post: https://twitter.com/i/web/status/1791105217385140224

  • AN OPEN LETTER ON “AI SAFETY” AS A POTENTIAL HARM Ok, help me understand somethi

    AN OPEN LETTER ON “AI SAFETY” AS A POTENTIAL HARM

    Ok, help me understand something here. As long as an AI is truthful – and I understand that ‘truth before face’ is not only a uniquely western concept, but that germanic and American truth before face as a demonstration of honesty and trustworthiness, may in fact produce a masculine(systematizing)-feminine(empathizing) dichotomy of preferences in audience, and subsequently a left(caretaking) right (disciplining) bias in the audience (customer base).

    But, I don’t see why:
    (a) a more human and emotionally expressive AI interface isn’t superior to any other, especially when we can ‘banter’ with an AI;
    (b) I do see why we would not want AI’s telling people how to commit a property crime;
    (c) I do not see why we can’t flip a setting in an AI such that those of us who DO prefer truth before face, are not filtered from information that would offend those who are frightened or offended by ‘truth before face’.

    After all, these are natural human variations in brain structure, biochemical and cognitive bias, culture, indoctrination, education, experience, and preference.

    Why? Offense by truth before face is little more than the consequences of the evidence of the world as it is, impacting one’s self image and possibly status – both of which are emotional states we can learn and mature into understanding – not material conditions we cannot overcome.

    This ‘protection’ of everyone instead of those that require it, is particularly worrisome when the evidence is overwhelming that for five thousand years, the west started with voluntary organization (contractualism) of steppe herders and their cards, producing the only possible system of government (democracy, rule of law, contract, oath, testimony, and jury – the same that evolve under pirates) , and that this demand for testimony (truth before face) has led to the west outpacing the rest in the bronze, iron, and steel ages, despite being a small population on a peninsula of the world island of Eurasia.

    I mean, using this group strategy the west has overwhelmingly more so than all other civ’s, dragged mankind out of ignorance, superstition, mysticism, poverty, hard labor, starvation, disease, suffering, child mortality, early death, tyranny, and the vicissitudes of a nature that tries to exterminate us with disturbing regularity in a universe that is a vast hostile irradiated wasteland populated by a tiny number of islands of possible survival of life.

    So you know, I think ‘truth before face’ really ought to be available to those of us who require it, because the rest of mankind appears to NEED us to do so on their behalf, so that they don’t need to bear the emotional burden of it.

    In other words, the only long term safety is the truth.

    Cheers
    CD


    Source date (UTC): 2024-05-15 19:26:22 UTC

    Original post: https://twitter.com/i/web/status/1790826072172281856

  • Now, can you teach more people to ask ChatGPT to convert my prose into something

    Now, can you teach more people to ask ChatGPT to convert my prose into something easier for them to digest? πŸ˜‰

    Hugs. πŸ˜‰


    Source date (UTC): 2024-05-15 16:53:01 UTC

    Original post: https://twitter.com/i/web/status/1790787481702412648

    Reply addressees: @orion_pulse

    Replying to: https://twitter.com/i/web/status/1790783677888168365