I am sometimes asked if I am concerned/restrict how our QAI SPECTRE™ super Turing analog neural computing synthesis technology is applied. It’s not a company relevant question, no more so than if Mentor, Cadence or Keysight were asked these kind of questions about their EDA tools. These kind of questions are all covered by existing import/export laws. A company cannot restrict who it sells its product to except as covered by existing law. More specifically, Apple can’t say, “I am not going to sell you an iPhone if you are not going to call your mom weekly”, for example. This is much different for companies that build application specific products with questionable or ambiguous primary or dual use since this gets into moral issues and social responsibility, (Google Dragonfly, which Google halted was an example), but for purely application neutral technology existing laws are the fundamental constraints a company can lawfully manage. #qaispectre #qai #brainchip
Adding Verilog-A behavioral synthesis and modeling to QAI SPECTRE™ super Turing BRAIN CHIP design, simulation and VLSI MCM neural computing synthesis system. Currently supports GDS2 and we are implementing OASIS synthesis soon. Goal is to eliminate need for third party layout, EM and simulation from QAI SPECTRE™ BRAIN CHIP design flow so that QAI SPECTRE™ synthesis is full flow through to photomask generation tapeout.
QAI SPECTRE™ super Turing analog recurrent neural network machines exceed all P-Space reducible (Turing limited) digital and quantum computers because of their ability to compute directly with irrational numbers (all P-Space reducible machines are restricted to rational numbers, quantum computers or otherwise). This enables QAI SPECTRE™ unlimited degrees of 3 THz+ per neuron problem solving freedom. QAI SPECTRE™ analog neural networks are dynamically self-organize using wireless semiconductor ultra high virtual fan in and fan out multi-level dynamic interconnect, enabling neuron instant response at sub millimeter wave frequencies to solve NP-Hard problems in linear time. We at QAI challenge any quantum computer company to demonstrate that QAI SPECTRE™ obsoletes any form of P-Space reducible quantum machines including any quantum parallel processor. QAI SPECTRE™ intends to dominate all digital and quantum computers using our QAI SPECTRE™ 4D on micro cooled diamond multi-chip modules, which are radiation hardened, self-healing, EHF, GaAs/GaN, analog and self-organizing.
(This is uncut for now, so skip thru.)
Lars Wood participates here in Risk Roundup to discuss "Machine Learning Bias: An Existential Risk."
A single cell amoeba solves the NP-Complete TSP in polynomial time since it implements analog super Turing processing as described in the Siegelmann-Sontag 1993 thesis upon which the QAI SPECTRE™ super Turing neural machines are based. Amazingly the researchers that conducted this study admit they have no idea why this is possible. It’s truly amazing how people blindly follow both digital computation and quantum computing (both P-Space reducible so both Turing machine architectures) because they think there is no other way. #QAI #QAISPECTRE Full paper is here:
Completed the UMS Foundry Course in France. Intensive three day course on the UMS end to end processes. This included SPECTRE technology discussions with UMS senior staff.
Dr Siegelmann discusses AI “Life Long Learning.”