judgment : tramp the halls of MWC 2023 , I get the chance to go manpower - on with Qualcomm ’s reference AR headset design . While the hardware is very much still in development and focused principally at manufacturer and developers , the demonstration introduced something I ’d never consider before : leveraging the clear AI program ChatGPT .
Qualcomm wo n’t really be producing AR headsets for consumers , but the society has resign a few reference AR headset design over the retiring couple of year that essentially serve as a pattern for other companies to make their own AR headset .
They ’re tardily have more heavyset and less like thebulky Microsoft HoloLens , but they essentially work in the same path , augmenting the real populace with practical element .
Manufacturers are already starting to leverage this technical school – theNreal Airis a good example of an early pair of AR glasses – but there ’s still a longsighted means to go before we ’re in a in full augmented future with slimline AR glasses . The headset I used , for good example , still requires a wired connection to a smartphone to process all the data from the headset .
That has change with Qualcomm’slatest AR Viewer Conceptpowered by the Snapdragon XR2 chipset that usesWi - Fi 7for ace rapid connectivity – Qualcomm claim less than 10ms latency – but due to the out-and-out amount of wireless equipment at the show , it was n’t possible to show it .
Still , I was capable to get some prison term with the more or less old Snapdragon XR1 reference headset , and I was presented with a Modern real - fourth dimension version and transcription service powered exclusively by ChatGPT .
ChatGPTis the red-hot topic in technical school at the moment , with the unfastened AI platform ostensibly capable to do anything from generate origin of code to speed up up app maturation to helping minor do prep . There ’s even a level about ChatGPTpassing step 1 of the Medical Licensing examination in the US , showing just how impressive the hardware is .
With that in brain , ChatGPT seems to be the idealistic prospect for veridical - time transcription and displacement with a focus on instinctive language modelling .
The headset used a dedicated microphone to clean up sound from the ( very loud ) demo orbit and nigh - absolutely canned what was tell on the trash , and it ’s voguish enough to detect multiple talker and split the transcription between them , with speaker 0 , speaker 1 and so on .
There is still a svelte lag while the API does its affair , but within a 2nd or two , I was seeing what was effectively live subtitle of my three - style conversation .
That might sound a bit gimmicky at a first glance , but it could be a real game - auto-changer not only for translation – imagine being able to talk to someone in a dissimilar voice communication using only a pair of glass – but arguably more importantly , a substantial boon for those that are hard of sense of hearing . The ability to effectively transcribe in literal - time could be freeing for those that are hard of audience , especially if discourse with someone that is n’t facile in signaling linguistic process .
Of course , there ’s still a way to go until we ’re all wearing AR headsets with ChatGPT - enabled transcription and displacement help , but it ’s a tantalising preview of what we should really carry to see run into the market in the next few years .
