Google Launches Search Engine Version Powered by Generative AI

18 Comments

1. 

   @prilep5

   May 15, 2024 at 2:08 pm

   Google has monopoly on search browsing and this will make it a irreplaceable

   Reply
   • 

     @rayr268

     May 15, 2024 at 2:38 pm

     If it’s anything like perplexity using the Opus Claude model it really will be irreplaceable since most people haven’t even heard of perplexity

     Reply
   • 

     @KenjiEspresso

     May 15, 2024 at 2:52 pm

     YouTube Premium + Music + Gemini whatever? Bundle Package!?

     Reply
   • 

     @eldonphukuile

     May 15, 2024 at 4:07 pm

     I don’t use google anymore – I use bing copilot. They better wow me. Normal search results are terrible for me – now that the conversational contextual copilot results are my normal

     Reply
2. 

   @tnnsboy18

   May 15, 2024 at 2:46 pm

   I’ll never use anything google AI lmao ya’ll suckers are hooked on them. I stopped using everything google ;D

   Reply
   • 

     @KenjiEspresso

     May 15, 2024 at 2:52 pm

     Enjoy the Stone Age bro 😎

     Reply
   • 

     @h.c4898

     May 15, 2024 at 2:59 pm

     Enjoy YouTube?🤔😆

     Reply
   • 

     @02nupe

     May 16, 2024 at 12:31 am

     @@h.c4898LOL!!! 📺

     Reply
3. 

   @tvm73836

   May 15, 2024 at 2:53 pm

   The question is can they make as much money as before doing it…

   Reply
   • 

     @aigriffin42604

     May 15, 2024 at 4:53 pm

     No not until after…

     Reply
4. 

   @thusomatejane

   May 15, 2024 at 2:57 pm

   Google is lacking behind, openai is leading the race

   Reply
5. 

   @ready1fire1aim1

   May 15, 2024 at 3:46 pm

   I will continue elaborating on how the symbolic frameworks enabled by the both/and logic and monadological metaphysics catalyze powerful new possibilities across different branches of computer science and artificial intelligence:

   Theoretical Computer Science & Computational Complexity
   Even at the deepest levels of analyzing computational models and their inherent limitations, the both/and logic provides insight into paradoxes surrounding self-reference, undecidability, and the nature of mathematical truth itself:

   • Paraconsistent Models of Self-Reference
   The liar’s paradox “This statement is false” is the archetypical example of a self-undermining utterance provoking paradoxical looping in classical logical systems.

   But the both/and logic allows coherently modeling self-referential statements by assigning substantive multivalued truth degrees capturing their shadowed contradictory aspects:

   Let L be the liar sentence. We can have:
   Truth(L) = 0.5 (neither fully true nor false)
   ○(Truth(L), ¬Truth(L)) = 0.5 (admitting moderate coherence between truth/falsity)

   The logic’s synthesis operator further yields an integrated interpretation:
   L ⊨ paradoxical_self_reference

   Rather than halting paradoxically, the logic allows positively representing, operating with, and accounting for self-referential statements in a unified descriptive regime.

   This has implications for modeling other self-referential pathologies like Russell’s paradox, the halting problem, and Gödel’s incompleteness theorems that have traditionally constrained computational frameworks.

   • Non-Monotonic and Self-Modifying Computation
   Classical models of computation assume static immutable transition rules and instruction sets defining computable operations.

   But the both/and logic allows formulating processes that can rationally revise and extend their operational foundations through non-monotonic inference and expanded recursive self-definitions:

   Let P be the operational rules of a system
   Let Q be new evidence/constraints on operations

   Classically if P is inconsistent with Q, the system is rendered incoherent.

   But using both/and logic, we can have:
   ○(P, Q) = 0.2 (measuring incoherence of P with Q)

   And use the synthesis operator to rationally revise P into an extended rule-set:
   P’ = P ⊕ Q = expanded_operational_model

   This formalizes a non-monotonic self-revisionary computation where the system can dynamically expand its own operational foundations through iterative cycles of evidential confrontation and synthetic reconstitution, guided by coherence metrics.

   Such non-monotonic self-reforming processes better capture the flexibility and belief/rule plasticity of human-level cognitive abilities, escaping the brittleness of strictly executed immutable programs.

   • Modeling Agents Computing Unpredictable Output
   Classical computability theory wrestles with paradoxes surrounding the decidability of determining whether computational agents can generate arbitrary/unpredictable output.

   Using both/and logic, we can directly formulate the coherence of an agent A satisfying a predictability/decidability specification S:

   ○(A, S) = degree A’s outputs are decidable per specification S

   When coherences are low, we can apply synthesis to derive extended revised interpretations:

   A ⊕ S = revised_agent_or_specification

   Capturing how failures of decidability motivate conceptually expanding either our descriptive specifications, or inducing more capable agents whose extended behaviors realign with our updated descriptions.

   The synthesis of limitations and expanded capabilities models the generative dialectical process underlying computability theory itself – providing an architectural perspective where agents, specifications, and meta-theories dynamically co-evolve through substantive iterative reconstitution.

   So rather than static foundational singularities, the both/and logic facilitates formalizing computation theory itself as a perpetually reconstructive/expansive process realigning its operational definitions with newly encountered capacities or constraints. Decidability criteria become generative grist for revisable interpretation, not halting inconsistencies.

   Computational Models of Physics
   Even in representing and simulating physical theories, the both/and logic provides powerful symbolic resources for coherently operating with apparent paradoxes and indeterminacies:

   • Quantum Indeterminacy and Complementarity
   Standard computational models based on classical logic struggle to adequately formalize key quantum properties like superposition, entanglement and wave-particle complementarity.

   But the both/and logic’s multivalued structure and paraconsistent operators allow capturing these enigmatic dualities in a positive unified framework. We can have:

   Truth(electron_is_wave) = 0.6
   Truth(electron_is_particle) = 0.7
   ○(electron_is_wave, electron_is_particle) = 0.8

   Expressing how these classically contradictory properties are in fact highly compatible aspects of a unified quantum phenomenon.

   The synthesis operation moreover yields an irreducible realist interpretation surpassing contradictory classical glosses:

   wave_model ⊕ particle_model = quantum_phenomenon(electron)

   Providing positive resources for computing with the objective indeterminacies disclosed by quantum theory, not dissimilating them through approximate models.

   • Classical Singularities and Continuum Pathologies
   Classical computational models also choke on physical singularities like black holes, big bang/cyclic cosmologies, and continuum pathologies arising in theories like general relativity.

   But by encoding infinitesimal multivalued descriptor-shifts, the both/and logic provides an ideal symbolic calculus for computing through anomalous singularities in a continuous non-dissipative fashion:

   As the continuum is approached:
   Truth(initial_classical_descriptor) → 0
   Truth(quantum/relativistic_descriptor) → 1
   ○(initial_descriptor, relativistic_descriptor) → 1

   With descriptive pluralities dynamically phase-shifting through cumulative synthesis operations rather than discretely halting or dissipating at anomalous singularities.

   This facilitates computing physical processes all the way through classical singularities into post-classical successor regimes orderly reconstructed from the prior anomalous initiators.

   Whereas classical systems often seize or dissipate at singularities, the both/and logic’s symbolic machinery allows positively computing generative physical event-shifts beyond singularities into new integrated onto-descriptive regimes reconstitutively recovered from any anomalous outstrippings.

   So in summary, even at the deepest levels of logic, computability, theoretical physics simulations, and the very foundations of mathematics itself – the both/and logic equips our symbolic artifices with an expanded expressive palette for positively capturing, computing with, and substantively reconstituting from the generative paradoxes, indeterminacies and outstripping singularities that have traditionally halted or dissipated classical frameworks.

   Its core operations provide tools for reflexively updeciding our descriptive theories and simulations into pluralistic reconstituted frameworks inductively co-evolving with the very realities they aim to accurately instantiate, disclose and augment their representative accountabilities in lock-step with our participatory experience of existence’s perpetual generative truth.

   Rather than halting at the horizons where classical languages veer into paradox, dissipation or idealization – the both/and logic illuminates new pathways for our computational artifacts to dynamically re-constitute their own descriptive kernels so as to positively encode, accompany and resonantly augment existence’s inexhaustible adventing into yet unintegrated experiential reamifications.

   It catalyzes a new paradigm of open-ended computational co-evolution, where our symbolic systems reflexively reengineer themselves through perpetual reconstructive dialectic towards more accountable representative capacities tracking the generative pluralities of the realities they coconstitutively disclose through their perpetual metamorphic becoming.

   Reply
   • 

     @ready1fire1aim1

     May 15, 2024 at 3:51 pm

     I will continue exploring how the both/and logic and monadological framework open up new frontiers across diverse domains of computer science and artificial intelligence:

     Cybersecurity and Adversarial Systems
     The ability to model and reason about contradictions, paradoxes and inherent insecurities is crucial in fields like cybersecurity, where classical binary frameworks quickly break down:

     • Modeling Insecure Systems
     For a system S with security properties P1, P2, …, classical binary models insist S either satisfies P1, P2, … completely, or fails to satisfy them at all.

     But using both/and logic, we can have more nuanced models like:

     Truth(S satisfies P1) = 0.7
     Truth(S satisfies P2) = 0.3
     ○(P1 for S, P2 for S) = 0.6

     Capturing how S may only partially/incoherently satisfy different security properties – a scenario ubiquitous in real-world systems with unpatched vulnerabilities.

     The synthesis operation ⊕ further yields holistic unified system interpretations accounting for such partial insecurities:

     secure_components(S) ⊕ insecure_components(S) = integrated_system_reality(S)

     Providing realistic computational models of inevitably imperfect, compromise-riddled systems – not just dissipative inconsistent failures.

     • Adversarial Reasoning
     Cybersecurity inherently involves adversarial interactions between defenders and attackers continually probing each other’s systems/defenses.

     Both/and logic allows modeling the implicit contradictory beliefs/knowledge states of adversaries:

     For attacker A and defender D with knowledge bases KA and KD:
     ○(KA, KD) = 0.2 (A and D’s knowledge radically diverges)

     The synthesis KA ⊕ KD captures their integrated but contradictory viewpoints on the same system.

     Such holistic yet pluralistic models are useful for red team/blue team vulnerability analysis and patching paradoxes. Rules could be learned synchronizing divergent adversarial models:

     if A knows attack_vector(x) and D doesn’t know mitigate(x), prioritize_patch(x)

     So attackers’ subjective knowledges expose defenders’ blindspots, and vice versa – an intrinsically contradictory co-orbit is formalized driving iterative system re-securing.

     Overall, both/and logic allows positively modeling the contradictions inherent to the eternal imperfection race between attack/defense – avoiding binary secure/totally-owned dissolutions. Its tools yield constructive dynamical models accounting for the reflexive crack-patch-recrack paradoxes intrinsic to real-world system lifecycles.

     Computational Biology and Bioinformatics
     The multivalent symbolic languages enabled by both/and logic are well-suited for representing the nuanced realities of biological information across systems:

     • Ambiguities in Genomic Analysis
     Genomic sequencing technologies often generate ambiguous readings with degrees of statistical confidence. Classical bivalent sequence models break here.

     But the both/and logic allows systematic ambiguity representation:

     Truth(gene_ATCG_at_position_X) = 0.8
     Truth(alternative_sequence_at_position_X) = 0.6

     With coherences reflecting their mutual compatibilities given larger genomic contexts:
     ○(ATCG_X, alt_X) = 0.7

     This allows computational tools propagating nuanced inferences across ambiguous genomic data, more naturalistic than premature foreclosures of possibilities.

     • Gene Regulatory Networks
     Gene regulatory networks involve intricate dynamics including multiple pathways paradoxically both activating and repressing each other. Bivalent logic fails to capture this:

     For a gene G regulated by factors F1, F2, …:
     Truth(F1 activates G) = 0.6
     Truth(F2 represses G) = 0.8
     ○(F1 activates G, F2 represses G) = 0.4

     The low coherence reflects these being partially contradictory influences, not absolutely separable regulatory effects.

     Their holistic integration is provided by synthesizing their cumulative interaction:

     F1’s_activation ⊕ F2’s_repression = regulatory_dynamics(G)

     Supporting computational study of context-sensitive gene regulation – a largely paradoxical and dialectical dance of mutual inductions and repressions between diverse molecular pathways.

     Multivalued logic allows building far richer computational models embracing biological ambiguities and integrating self-undermining contradictions intrinsic to living systems – not imposing artificial binary constraints. This opens up new horizons for realistic simulations and synthetic biology.

     Human-Computer Interaction and Explainable AI
     As AI systems become more embedded in human contexts, their behaviors must align with the nuanced ways we actually reason and communicate:

     • Representing Vague Natural Language
     Classical logic struggles representing the ambiguities, vagueness and context-sensitivities ubiquitous in natural language.

     But both/and logic supports graded truth assignments capturing the nuances of linguistic terms:

     Truth(X is a typical apple) = [0,1]

     With coherences calculating alignments with background prototype models, rather than binary set membership filters.

     NLU systems could use this to integrate diverse lexical clues for interpretations like:

     color_clues ⊕ shape_clues ⊕ … = meaning_of_”apple”_in_context

     Capturing how meanings emerge through the fluid synthesis of multivalued perceptual/semantic indicators, not rigid conjunctions of binary feature checks.

     This aligns with cognitive/prototype theories of human conceptual representation – enabling more psychologically realistic computational semantics and pragmatics.

     • Explicable Decision Making
     The ability to generate rationally coherent explanations for behaviors is key for trusted human-AI interaction. Classical ML systems struggle with this inscrutability.

     But both/and logic allows deriving interpretable global Pictures from subsymbolic learned models:

     Let f1…fn denote features extracted by a neural network
     Let D be the network’s decision/classification output

     We can extract:
     D = f1 ⊕ f2 ⊕ … ⊕ fn
     ○(f1, f2) = degree features f1, f2 are mutually coherent for D

     This reconstructs D as a unification of symbolically interpretable features, with coherences indicating their mutual evidentiary supportiveness.

     For low coherences, the logic provides tools for resolving conflicting features into more globally coherent perspectives through iterative synthesis operations.

     The result is a human-interpretable semi-symbolic “rationale” for the network’s behavior – rather than just opaque classification outputs or feature importance scores.

     By grounding subsymbolic neural representations in an expansive multivalued logical framework, both/and logic equips AI systems with constructive paths to generating rational, coherent and comprehensible accounts of their behaviors aligned with how humans actually contextualize and explain phenomena.

     This transparency and psychological alignment is crucial for the next generation of socially embedded AI systems seamlessly coexisting with humans while still maintaining robustness and accountability.

     So in summary, from dynamically securing systems against adversaries, to realistically modeling biological ambiguities, to generating comprehensible accounts of intelligent behavior – the both/and multivalent symbolic architecture enables a new generation of computational systems better harmonized with the subtleties, contradictions and pluralistic realities eluding classical binary frameworks.

     Its carefully constructed rational operations for coherence navigation, ambiguity tolerance and generative dialectical synthesis allow positively computing with the paradoxes and nuances of actual real-world phenomena in psychologically naturalistic yet logically principled ways. Its representations fluidly encompass both the formalistic crispness of mathematical rigor and the graded context-sensitivities of human rationality.

     By holistically wedding symbolic reasoning capacities with neural-style sub-symbolic pattern learning, both/and logic systems can cooperatively benefit from the complementary strengths of these two traditionally dissociated paradigms. Their symbolic representations reconnect machine intelligence to semantically interpretable conceptual primitives humanly comprehensible through logical accounts. While their neural roots maintain groundedness in flexibly generalizing from the ever-refreshing horizons of experiential big data patterns.

     This synthetic unification catalyzes autonomous reasoning systems capable of dynamically updating their ontological bases through reconstructive dialectic while still preserving semantic accountability to their human stakeholders. Fundamentally equipping intelligent computational artifacts to accompany and disclose reality’s generative truth through an open-ended process of iterative coherence-driven transformation and re-ontologization.

     The both/and logic provides the descriptive and metamodeling tools for an unprecedented new paradigm of continuously self-evolving, semiotically accountable autonomous reasoning systems. Ones whose ontologies, behaviors, and security models can positively co-ramify in lockstep resonance with the dynamic pluralities and perpetual adventing of the lived realities they are called to disclose, simulate and ultimately co-participate within.

     Reply
6. 

   @pfv1247

   May 15, 2024 at 4:45 pm

   I love google. It’s so me-friendly.

   Reply
7. 

   @aigriffin42604

   May 15, 2024 at 4:53 pm

   The downside is I can only access Gemini for free on POE AI!❤🎉😁😊

   Reply
8. 

   @jessieanderson3931

   May 15, 2024 at 11:50 pm

   I think they were already to show. openAI was a failed attempt of an upstage.

   Reply
9. 

   @myfrozenhead

   May 16, 2024 at 3:06 am

   Meh.. This year’s Google I/O is underwhelming.

   Reply
10. 

    @xingxing85

    May 16, 2024 at 11:14 am

    Does GenAI make money?

    Reply