Often in marketing we talk about switching costs (what does it cost to ditch the current method or arrangement and switch to a new approach) and gold plating compared with adopting a "good-enough" solution.
In the case of power efficiency, sometimes its cheaper to find more power than it is to find more efficiency particularly if that means developing a whole new system on a brand new platform. And remembering that engineers that have mature AI capability are thin on the ground let alone those that are familiar with Akida.
Further, there are many examples of industries that put-up with antiquated technology when you would think they had every reason not to...
- The stethoscope is still a primary medical diagnostic tool as is the mercury tube Sphygmomanometer (measures blood pressure).
- Many banks and finance companies still run monolithic mainframe core systems programmed in COBOL. They work, And they are given shiny new front ends by interfacing them to modern computer systems that provide graphical user interfaces and API's that allow internet banking, talk to ATMs, allow data analytics etc.
- The entire global airline reservation system at it's core uses a computer system written in the 1960's. But, enhanced through middle-ware that talks to it.
- The drastic rise in unemployment due to COVID-19 revealed the US unemployment systems’ inability to handle a large number of cases.Most states use unemployment systems that are over 40 years old and run on COBOL language.
- In the USA In retail, corporations spend 58% of their IT budget on maintaining legacy systems. The main reason why businesses use legacy software is that the sales terminals use Intel 286 computers.
- The US Airforce still flies B52 Bombers (1950's technology with upgrades) - because for many missions they work perfectly well.
- Many Fortune 500 companies run core systems using EXCEL spreadsheets.
- All Doctors and Pharmacies in Australia still use FAX machines as their primary communication system (as my wife recently found out when a specialist's secretary rang to say they hadn't received the referral and could she FAX it through? - she had to go to a post office).
I could go on.
But, of course that doesn't mean that there is no market for new technology. While all these legacy systems have remained stagnant, clearly the world has changed massively around them.
AI at the Edge and anywhere else is here, new applications are being explored and developed. But the key features of this transition will be...
- Slow to begin with (Hype far exceeds reality)
- A period of disappointment - "we was conned"
- A sudden explosion of gob-smacking and game changing new tech - A STEP change.
- Followed by a long period of massive rollout and incremental improvement.
Akida (and what follows) will be in there somewhere.
Just don't know when its going to happen.
IMHO
(20min delay)