3 ms·
I really can't understand the hatred directed towards Magic Leap. All we know at this point in time is a) those who have experienced the technology are impresse
by decision_tree 9y ago
I really can't understand the hatred directed towards Magic Leap. All we know at this point in time is a) those who have experienced the technology are impressed (but obviously bound to NDA) and b) ML has raised near-unprecedented amounts of money for a company with no product on sale (2.3MM by CB[1]). To think Google and Alibaba, not to mention other big funders, are so wrong in their large contributions is serious - such a mistake is historic. I'm not suggesting they will deliver, only that all we can say is we don't know.
[1]https://www.crunchbase.com/organization/magic-leap https://www.crunchbase.com/organization/magic-leap
- wolfcode 9y agoMaybe you would better understand if you read technical source like: https://www.kguttag.com/2018/02/22/magic-leap-one-tunnel-vision/ https://www.kguttag.com/2018/02/22/magic-leap-one-tunnel-vis... They sell dreams. Unfortunately they got social proof - "quickly, other already gave us a lot of money, you should too!". There is already existing HoloLens tech which you can already buy.
- edanm 9y agoThis assumes a very high level of naivety by the people funding the company, who at least theoretically, should be, or have access to, technical experts that can study the technology for themselves. We're not talking about people raising money from the general public or using heavy-handed sales tactics. We're talking about billionaires funding this thing - I'm not going to cry if they lose money on a bad investment (unless there is outright fraud involved).
- wolfcode 9y agoIt's was not very high level of naivety to invest in ML back then, they did impressive marketing job and innovation most often happens in smaller companies - it is far more easy to acquire company shares than try to build a team - hard to hire experts, build labs and finally this requires time. Note that ARKit/ARCore are steps required for this kind of tech - so at least Google and Apple are building expertise on AR/MR software technology, but the hardware part is still left for others to figure out. Today except Microsoft there is Intel using Sony's waveguide simple display and other headsets that are not giving magic experience by using more primitive tech. Anyway for me it is sad they still invest in ML only for this reason that supporting more players would greatly improve chances that some company would figure out better solutions and ML is stuck in local minima to keep most of their story and promises.