A reminder to all: There is a MeTTa-on-top-of-AtomSpace proof-of-concept at https://github.com/opencog/atomspace-metta that needs an active developer to extend it. The general idea here is that the AtomSpace is already distributed/decentralized, its already fast, and its stable/debugged, and it has python bindings. I think it remains viable, if that's the direction you want to go. But what do I know.
-- Linas On Fri, Sep 15, 2023 at 1:14 PM Ivan V. <[email protected]> wrote: > Thank you, Ben, for taking the time to answer. > > Here is the current market situation: > > -- as a language for scripting neural-symbolic, symbolic, >> evolutionary, and other sorts of hybrid cognitive architectures and >> AGI systems > > > Google: agi programming language > <https://www.google.com/search?q=agi+programming+language> > > -- as a smart contract language > > > Google: smart contract programming language > <https://www.google.com/search?q=smart+contract+programming+language> > > -- as a framework for creating DSLs, especially those that need to >> operate in a secure and/or decentralized setting (e.g. domain-specific >> smart contract languages) > > > Google: framework for creating dsl systems > <https://www.google.com/search?q=framework+for+creating+dsl+systems> > > - - - > > Finally, it may be worth of throwing a look at these interesting > frameworks - they all, more or less, somewhat resemble a kind of applied > typed lambda calculus: > > Google: framework for defining formal systems > <https://www.google.com/search?q=framework+for+defining+formal+systems> > > - - - > > Interesting times indeed. > > ivan > > pet, 15. ruj 2023. u 18:47 nugi nugroho <[email protected]> napisao je: > >> Hi Ben, from the last agi conference, there was a mention of a >> hyperdimensional architecture that can potentially rival ANN. Considering >> the development of the hypervector chip, metta language and other chips >> designed for this agi project, how long do you expect (assuming there are >> few thousands people running that chips in the nunet >> network coordinatively) when we will achieve performance equivalent of gpt3 >> by using hyper vector model or something like categorial grammar or link >> grammar combined with other AI model in MeTTa language to make a >> transparent and ethical AI. This project is one of the reasons I am >> majoring in CS. Will I have time to master the AGI field and contribute in >> this field? >> >> Pada tanggal Jum, 15 Sep 2023 pukul 22.47 Ben Goertzel < >> [email protected]> menulis: >> >>> We haven't really been thinking about marketing or developing MeTTa as >>> a broad-scope programming language, rather we've been thinking about >>> it >>> >>> -- as a language for scripting neural-symbolic, symbolic, >>> evolutionary, and other sorts of hybrid cognitive architectures and >>> AGI systems >>> >>> -- as a smart contract language >>> >>> -- as a framework for creating DSLs, especially those that need to >>> operate in a secure and/or decentralized setting (e.g. domain-specific >>> smart contract languages) >>> >>> Integration w/ other already-popular and "easier" languages like >>> python and Rust is being prioritized to a significant degree, at least >>> at this stage, rather than development of MeTTa as mostly a standalone >>> software development framework... >>> >>> We are also already ofc looking at use of LLMs for mapping natural >>> language into MeTTa (initially in the "AI-DSL" context of using NL to >>> describe tasks and having an LLM translate this into MeTTa code for >>> combining together multiple existing software agents (on >>> SingularityNET platform) to execute these tasks...). The ability to >>> go from NL to MeTTa may to some extent overcome the obstacle of MeTTa >>> being too abstract for most programmers to handle. We are entering >>> into an era where the number of software developers who need to >>> actually. master programming language syntax will rapidly become >>> smaller and smaller... >>> >>> BTW unlike with most weird functional/logical programming languages we >>> are attempting to be aggressive about scalability from the start. >>> The collaboration w/ Greg Meredith's group is proving potentially very >>> interesting in terms of getting to a high-speed compilation and >>> execution process for MeTTa... we are looking at compiling MeTTa to >>> Rholang and getting a major speedup that way (largely due to efficient >>> use of multiple cores and threads), and then mapping rholang to >>> hypervectors which can be very efficiently manipulated on the GSI APU >>> chip, and getting major additional speedup that way. There is also >>> some work underway on a custom processor-in-memory chip for MeTTa >>> pattern matching but that is still at the simulation-modeling stage... >>> >>> Rough timeline is we hope to have Hyperon alpha version by March 2024 >>> or so.... this won't include all this fancy Rholang--based >>> acceleration but potentially some of it... it should also include >>> integration of the local Hyperon stuff w/ Distributed Atomspace >>> (currently back-ending on Mongo+Redis) ... Once the alpha is launched >>> it will be the time to start aggressively building an OSS dev (and >>> research) community around Hyperon, I think... >>> >>> Experimenting with some genomics-oriented reasoning, and dialogue >>> systems for robots + avatars, as among the early experimental >>> use-cases for all this... >>> >>> Fun times, eh? ;-) >>> >>> -- Ben >>> >>> On Fri, Sep 15, 2023 at 7:23 AM Ivan V. <[email protected]> wrote: >>> > >>> > Hi Adam, >>> > >>> > As I understand, Verse lang is something like MeTTa without AtomSpace. >>> Considering that difference, probably an integration of lang and database >>> may bring some benefits, as with performance, so with integration >>> simplicity presented to end users. >>> > >>> > Probably the market that MeTTa can potentially cover is practically >>> anything related to computation. But also probably, that may not be the >>> best way to promote the product, as the agenda message may disperse too >>> much without the clear designation to this or that task. So, I believe that >>> an important question may be what piece of the existing market to claim >>> with MeTTa, or even what new kind of market to create and conquer. >>> > >>> > Regards, >>> > ivan >>> > >>> > pet, 15. ruj 2023. u 15:48 Adam Vandervorst < >>> [email protected]> napisao je: >>> >> >>> >> Hi Ivan, >>> >> >>> >> >>> >> From a non-deterministic distributed programming language with >>> unification angle, my guess would be the Verse programming language ( >>> https://simon.peytonjones.org/assets/pdfs/verse-conf.pdf). >>> >> >>> >> >>> >> Best, >>> >> Adam >>> >> >>> >> Op vr 15 sep. 2023 16:37 schreef Ivan V. <[email protected]>: >>> >>> >>> >>> Hi all, >>> >>> >>> >>> Will MeTTa have any competitors? If it will, who will they be, and >>> what will MeTTa do to outperform them? >>> >>> >>> >>> All well, >>> >>> ivan >>> >>> >>> >>> -- >>> >>> You received this message because you are subscribed to the Google >>> Groups "opencog" group. >>> >>> To unsubscribe from this group and stop receiving emails from it, >>> send an email to [email protected]. >>> >>> To view this discussion on the web visit >>> https://groups.google.com/d/msgid/opencog/9fda4668-5041-4a94-bdd2-b69441090614n%40googlegroups.com >>> . >>> >> >>> >> -- >>> >> You received this message because you are subscribed to the Google >>> Groups "opencog" group. >>> >> To unsubscribe from this group and stop receiving emails from it, >>> send an email to [email protected]. >>> >> To view this discussion on the web visit >>> https://groups.google.com/d/msgid/opencog/CAHxQD9i42M7P1_3Wxg4zWu3Hg-O3pHJWdCmuhyzhbynCt-m1%3Dw%40mail.gmail.com >>> . >>> > >>> > -- >>> > You received this message because you are subscribed to the Google >>> Groups "opencog" group. >>> > To unsubscribe from this group and stop receiving emails from it, send >>> an email to [email protected]. >>> > To view this discussion on the web visit >>> https://groups.google.com/d/msgid/opencog/CAB5%3Dj6WxXH2Ab7dduHj3q0F2vR56%2BJLQZvBNcHLbNCBAv0L6hw%40mail.gmail.com >>> . >>> >>> >>> >>> -- >>> Ben Goertzel, PhD >>> [email protected] >>> >>> "My humanity is a constant self-overcoming" -- Friedrich Nietzsche >>> >>> -- >>> You received this message because you are subscribed to the Google >>> Groups "opencog" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to [email protected]. >>> To view this discussion on the web visit >>> https://groups.google.com/d/msgid/opencog/CACYTDBeHJomQS3VynFdeUYXCrh%2BRG%2BzrE9%2BEGjt7LC-J55D7EQ%40mail.gmail.com >>> . >>> >> -- >> You received this message because you are subscribed to the Google Groups >> "opencog" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected]. >> To view this discussion on the web visit >> https://groups.google.com/d/msgid/opencog/CAJA72CdQicRhPNMXjyYk10Sc30d5ss51Z0t9N%3DA-36jzyuZmog%40mail.gmail.com >> <https://groups.google.com/d/msgid/opencog/CAJA72CdQicRhPNMXjyYk10Sc30d5ss51Z0t9N%3DA-36jzyuZmog%40mail.gmail.com?utm_medium=email&utm_source=footer> >> . >> > -- > You received this message because you are subscribed to the Google Groups > "opencog" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To view this discussion on the web visit > https://groups.google.com/d/msgid/opencog/CAB5%3Dj6Wfr5dmDLH3%3DO_rYQZNA81j4ZfPy%2B6AemcwE2%2B9rVL38g%40mail.gmail.com > <https://groups.google.com/d/msgid/opencog/CAB5%3Dj6Wfr5dmDLH3%3DO_rYQZNA81j4ZfPy%2B6AemcwE2%2B9rVL38g%40mail.gmail.com?utm_medium=email&utm_source=footer> > . > -- Patrick: Are they laughing at us? Sponge Bob: No, Patrick, they are laughing next to us. -- You received this message because you are subscribed to the Google Groups "opencog" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/opencog/CAHrUA37%3DCYHXzqYW2Ywq5xkHb2T0%3DBGrvOPHQ%3Dix4EiJK2czgA%40mail.gmail.com.
