uiux
Regular
"What I envision for this company is that if we proliferate the technology, prove it out even more in many, many industries and use cases, that we become the defacto standard. That is the vision I have for the long term. There is absolutely no reason we cannot be the predominant standard for edge AI in the coming years"
- Sean Herir, BrainChip Podcast
~3m55s
https://open.spotify.com/embed-podcast/episode/21BiwGFXUTnbBRU46WELBz
---
To look at some recent things that became "de facto" - I checked the wikipedia page and found a few relevant items that I think we could use as a basis for comparison.
The MP3 audio format
started as an alternative to WAV for Internet music distribution, then replaced it. It is now supported by the vast majority of music players, audio transport, audio storage and non-commercial media. WAV and MP3 are also "de jure ISO formats".
Was first created in 1993 by Adobe. Adobe internal standards were part of its software quality systems, but they were neither published nor coordinated by a standards body. With the Acrobat Reader program available for free, and continued support of the format, PDF eventually became the de facto standard for printable documents. In 2005, PDF/A became a de jure standard as ISO 19005-1:2005.[2] In 2008 Adobe's PDF 1.7 became ISO 32000-1:2008.
In BrainChips case, I would think that the primary feature that would lead it to becoming "de facto" would be the edge learning capabilities. In a BrainChip patent awarded last year, it discussed the neural libraries:
https://patents.google.com/patent/US20190012597A1/en
Method and A System for Creating Dynamic Neural Function Libraries
Abstract
A method for creating a dynamic neural function library that relates to Artificial Intelligence systems and devices is provided. Within a dynamic neural network (artificial intelligent device), a plurality of control values are autonomously generated during a learning process and thus stored in synaptic registers of the artificial intelligent device that represent a training model of a task or a function learned by the artificial intelligent device. Control Values include, but are not limited to, values that indicate the neurotransmitter level that is present in the synapse, the neurotransmitter type, the connectome, the neuromodulator sensitivity, and other synaptic, dendric delay and axonal delay parameters. These values form collectively a training model. Training models are stored in the dynamic neural function library of the artificial intelligent device. The artificial intelligent device copies the function library to an electronic data processing device memory that is reusable to train another artificial intelligent device.
What stands out to me is that during the patent application process, some Microsoft patents were cited, specifically this one:
https://patents.google.com/patent/US5247681A/en
Dynamic link libraries system and method
Abstract
A system and method for sharing software modules which are part of a computer program, and which have been previously loaded into a private area of main memory of a computer. The addresses of software modules which remain loaded in the private area after an execution of the computer program are kept track of. When a subsequent execution of the computer program requires a software module which has remained loaded in private memory, the address of the software module is obtained, and the computer program branches to that address.
if you have used Windows enough and received some BSOD you might recognise the reference DLL when reading the error screen.
What BrainChip has created is similar to DLLs, neural libraries that can be created, stored and shared. I understand this is the sort of innovation that will help push the technology into becoming "de facto". I don't think its entirely based on hardware or performance, but more the learning capabilities and the methodologies of sharing this knowledge amongst other artificially intelligent devices.
all in my opinion.
