XNOR.ai Terminates Wyze
Wyze has shocked the industry again - this time with a major setback. Just months ago, Wyze triumphantly added AI analytics to their ~$20 cameras, which IPVM testing showed worked better than Axis and Hikvision analytics.
Now, Wyze has announced XNOR.ai, their analytics provider, has terminated them, leaving Wyze scrambling.
Inside this note, we examine the challenges for Wyze and the opportunity for XNOR.ai, the startup that developed YOLO, to expand in video analytics.
Terminating ****
**** **** * ****** ***** ****** AI *********,**** ************* *** ******** **** **:
*** ***** ******** * *** ******* which ** *** *** ******* ****** into ****** ** *******. *** ** those ******* *******Xnor.ai ** ********* *** ***** ** *** ****** ******* ******, forcing our next firmware release to omit the Person Detection feature. [emphasis added]
**** ** ** ************* ********* ** (1) ****** ** ********* ***** **** issues *** (*) ****** ** **** enter **** ********* *** **** *** features ******* ********** ** *******.
2 **********
***** ******* **** *********, ** **** two ********** ****:
- **** ***** *** *** *** ***** that ****.** ******. **** *** ** was ********** *** **** **** ** inferred **** ******** **** ****.**, ****.** was *** ***** **** **** (** anything) *** *** **** **** *** that *** ******* **** ****.** *** was ****** ********* (*** ********** ********* at **** *** * ******* **** minimal ***** *********** ****** **** ****). So ****.** *** **** ****** **** to *** **** ** **** *** clearly ******* *********** ***** **** ****.**'* technology.
- ****.** *** **** ********* ****** ** do ****, ******* **** **** * larger ********, ******* ** ******** **** views **** ** * ********** *** wants ** ******* **** **** ****** access ** ****.**'* **********.
Benefit *** ****.**
****** ***, **** ** ****** ** be * ******* *** ****.** ** they **** **** **** ** ***** that ***** ********** ***** ** **** extremely ************, ***-**** ***** **** **** uses.
Problem *** ****
**** ******* ******** **** **** *** not ******* **** ** ** **** can ******* * ********** *** ******* on ***** ***-**** *******, ******:
***good **** is that when we do re-roll out this feature it will still be free for you - even if we end up using the cloud instead of edge-AI.
***? ******* **’** *** ** *** business ** ****-***-********* *** ***** *** we’ve **** * ******* ** *** that ****** ********* ***** ** * free *******.
** **** ** **** ***** ********* for ****, ** ** ***** ** cut **** ***** ******* ******* *******.
Wyze ****** ** ****
** **** **** **** *******, **** says **** *** ******** ***** *** AI ****:
** ******* *** ********** ** **** feature *** *** ********** ** ** going ******* *** *** *******, **’** already ****** * **** ** ******** AI ********* *********** ******* ** *** Person ********* *******.
***** ** ** ***** **** **** can ******* *** *** ** ** there ** ** ******* ****** ** publicly ********* ********* ** ***** ** in ******** ****** (********* **** ****.**). The ********* ** ** **** ***** models *** ********** ** *** **** of ************, ***-**** ***** **** *** in ****'* *********. *********, *** ******** cost ** ***** ** ** *** cloud *** **** ******** ** ******* could ** **** ********* *** ********* the **** ******** *****.
"Too **** ** ** ****"
**** **** ******** *** ****.** ******* AI, ***** ******* **** ***** **** was ** "************ **** ********** ** create *** **** ** ** **** products":
**********, **** *** **** ****.
Outlook ***** ******** *** ****
***** ****** ** **-***** **** *** literally *******, ******** ** *** ***** of *** *** ** ****, *** losing ** ** ************ *** * setback, **** ***** *** ** ********* combination ** ***** ***-**** *** ******* marketing **** **** ******** ** ***** its ********* **** *** ******** ******.
*******, *** **** ***** *** ******'* ******* ******* ******* ********* **** **** ** ****** ******* challenges ** ** *******.
Vote / ****
*** **** ********* ****
**** ********** ** ********* ** ***** ************, **** ***** ** ** *** case. *** ***** ***** **** ** limited *********** ** ** **** ***** price ***** **** *** **.
***, ********** *** * **** ***** them ** ****** ** ******.
** *** ****** ** ***************** *** ***** ** ********** **** ******* *** ****** ** how *******'* *** *** ********* ** marketing.
** ** ***** ********* ** **** / **** *** ******** *******, *.*., the ****** *** ***** ***** **** to *** **** ******* ***** ** are ** **** ** ** ** exact - *.*.,******* ****.**, ********* ** ****, ***** Surveillance ***** ********. ** **** ***, **** *** simply ******* *** **** ** **, but ***** ******** *** ** ****** as ** ********* **** ** *********, which ** ** *** *****.
** ** ***** ********* ** **** / **** *** ******** *******, *.*., the ****** *** ***** ***** **** to *** **** ******* ***** ** are ** **** ** ** ** exact...
***, *** **** ** *** **** has *** ** ******* **** *** “now **********” ******** :)
******** **** *** ****.
* ** ******* ** *** **** prompted ****.** ** **** **** ****. Seeing *** ***** ***** **** ** is, * **** **** *** **** landing ** ******** **** ****** * traditional / ****** ***.
**** ***********! * ****** ******** ***** can ****** *** ** ******* *** firmware **** ******* **** ****** *********. I **** * *** **** **** deployed ** **** *** ***** **** mostly ******* ** *** ****** ********* analytic. ** ***'* ** ***** *** them ** ****** * ***** ************* after *** ********...
** ***** ** ******** **** *** don't **** ** ******* ** *** patch **** ******* **** ****** *********, you *** ******* **** ** ****** firmware ******** *** **** ** ***** firmware.
* ******** ***** ******/*** ******* ****** which **** *** ** ****** ** whether *** *** ****** ******** ******, tech ******* **** **** *** *** stop *** ******* ***** **** ******* an *** ****** *** **** **** I ***** *** ******* **** **** me "******* ** **** ****" *** "can't ********* **** ** (********** *** upgrade) **** ****".
* ***** "****", *** ** **********, how *** *** **** **? ****'* market ** ********* ***** *** ****** cameras. ***, *** ********* **** ******* to **** *** ***** **** ****** up * *** ** *******, *** just **** *** **** ***** ****.**, I ***** **** ****** ** *** Wyze's **** *** *****. ** ** hadn't **** *** **** ****, * probably ***** **** ******* ** **** as **** **** ***** ****** *******, but **** ***** *** **** *** originally *********, *'** ********** **** ****** of ****** ********* **.
** *** ***** ***, **** **** set **** **** * ******, *** I ****** **** ** ******* ***** their ******** ***** ** ******* ***** cameras, *** **** **** **** ***** getting **** *** ***** ***** *****, they'll ** * ***** ** ** reckoned ****.
*** ********** ****
* ***** **'* **** ****** ** essentially ** **** - ****.** ****** more ***** **** **** **** **** were ******* ** *********** ***, ******* XNOR ****** *** ********* ** **** that **** *** ************* **** *** a ***** ** ***********, ** **** believes **** **** **** ***** *********.
**** ** * ******* ******* ** the ******* ***** ** ********* ******** - *** ***** **** *** "**** cost" ** ********* *** ************ ** venture *******. **** **'* ***** ********** by * *********** ******** - *** enough ***** (** **** * ****) being **** ** *** *********, *** as ** ******* ** **** *** adoption **** ** **** ** ********, figure *** ****** *****.
**** ******** ****** - ** **** point, ***** ******** ** ******* ********* through ******* *************** *** ***** *********** seems ** ** *** **** ****** successful **** ** *** (******* *** whole ******* ****** ******** **** ** the ****).
**** ******** ******
****: **** ** ******** **** * Avigilon **********. ** *****, * *** thinking *** ****’* *** **** ** a **** ********** **** ** ******** employee *** ** *** ***.
***** ** ** *** *** ******** while *** ***!
**.*% ** **** ********* ***'* **** the **********. **** ****** ***'* *** the ******, ** *********** **** **** had...
**** ****** ***'* *** *** ******, or *********** **** **** ***...
**** **** *** **** *****. **** Wyze ************ **** ********** ****** ********* was *** ** ***** ********'* *** complaints. ****** **** *** *********** **** out * ********** **** ***** **** as ****, * *********** ****** ** customers **** ****** *** ** ************.
******** * ***** ** **** ** a ****** ** ****. **.*% ***'* realize *** **** **** *** ** because **** ***'* ****** ** ****, yet.
** *********** **** **** ***...
*** ******** ********, ** *** *****, the "***'* **** ** **** ***'** doing *** ** ***, ****** ** of *** ***** *** **** **** from **" ********.
** *** **** ******* ******* *** consumer ***** ****** *******, ******** **-**% of ***** ******* ** ******* ** motion ********* ******* *** ************* ******.
* ******, **** * ******* ***** it *** **** ****** ** ********. One ***** ***** **** * ****** paying $** *** * ****** ***** not ****** **** ***** ******** ** long ** *** ******* ****** ********.
*** **** ** ********** *** ********* who *** *** ****** **** ******** generally ****** *** **** *** ***** have *********** ************ ** **** * product *****/****** **.
* ***** **** ********* *** **** synergy. ***** ********* ***** **** ***** on * ******** *****. **** **** Wyze *** ******** ********. **** ****** people ********* *** ****. *** * long **** ******** ***** *** *** work ***.
*******, * ***** **** **** ** just ** ***** **** *** *** trend:
** ** ****, ** ***** *** it's **** ** ***** ********.
** ** *** ***** ** **** major ******* ** *** ********. *****, IMO, *********** ******* **** **** ** polish ***** *********. **** **** **** to **** ** * ********.
* ******* ******* **** ******** ****** and ***** * ****** ** **** users *** ******* ****.**, **** **:
**'* ****** ***** **** **** *** claiming **** *** ****** **** ***** are ** *********** **** **** *** actually ****** **** / ******** ***** from **** ***********.
** **** * ****?
@*********** *** ******* ****** ***** **&* and ********....
*****************, ** **.
********* **** **** ********* ******* **** want ***. *** *** *****. **** like ***** ***** ********/ ** ******* out ***** ****** **** **** **** RMR ** ********** ******* *** *** really *** ** ****** **** ****.
*****, ******? **** ****** * ******* few, ** ***, **** ** ******** analytics ** ****** ** *** *** chipsets. *** *** *** ** **** them *** *** **** ** ******? It ** ****** ** ***** *** should *** **** ******* ******* * monthly ***.
*** *** *******'* **** *** *** RMR. **** ** ******** ***, **** fair ****.** **** ***** **** ***** of ****. ** ***** ***** ** develop ***** **'*, ********** ****** ******* and ******* *** *** ****. ** either **** * ***** ** *** RMR, ** *** *** ** *******. I ***** **** *** ****** **** lump *** ********, ** * ******* they ******* **** ****** ***** ****.
**** ** ******** ***
**** **** ******* ** ****** *** RMR * **** *** **** ***** '******** ****** *******' ******** ** $*.** *** ****** per *****.
**** ** *** *** **** **** focused ** ******* ** **** ***** as ******** ** *** ****** ******** price.
** ** *****, * ** *** defending **** ****, **** ********** **** role *** ***** ** ****'* ********.
************ ****** *** ** *******?
*** *** ** **** *** ****.** pulled ***** ******** **** **** - they **** ***** ******** ** ***** and ***** ******** ****'* **** ****** else **** ******/***** *** ********.
***** **** ****.**, ** ****-******* *** spin-out, *** ***** ** $**** ***** – ********
***, ****** *** *******!
**** **** *** ****.** ** ****** us **** **** **** ***** ******** out *** ******** **** *** **** is ** ********* ***** **** *** them.
** *** *** ** **** *** the ******** ** **** ****** ***** Apple **** **** **** *** *** own ******** *** *** ******* *** IP *******, *** ** **** ***.
* ***** ******** ******** *** ***** on * *** **** *** ***'* see **** ******** ** ** *******. Apple **** **** ** *** ********** (if **** **** *** **). ******* up * ***** **** ** **/****** Network ******* ** ***** *** $**** spend. **** ** *** **** ***** these ***** ** ******* *** ********** extraordinary ************.
* ***** **** **** ********* **** to ******* ******* ****** ***** ***** works **** ********** ***** ***** *******. It ******** ********* ** ****** ****** than **** ** ***** *** ********. This ** *** ** ************ ****** ***** *********: *****'* **** to **** **** ******** *** **** safe | *********
*****. ***** **** **** ** *** themselves *** *** ******** ***** **... Just **** **** ********* ******* * competitor *** *** **** ******** ***. What's *** *****...
******, * **** *** * **** later, **** ** ***** ******* ** replacing ****.**, ****** ******:
** ****-**** ***** ****, ***** *** whole ****.** ****** ** **** * BIG *** *** ******* ********* ******* into *** **-***** **** ********* ****. Now ** **** ** **** **.** we ***** ******** ***** * ***** university.
Here’s *** ****** ** **** **.
**** *****, ** ***** **** ** thousands ** *** ****** ** *** algorithm ** **** ** *** *** Detection ******* **’** ******* **. **** will **** ******* **** ******* ***** recognized ** ******. (** ***** ****’** Centaurs?)
**’** ****** **** ******* *.* ** our ******-***** ****** *********** *******. **** time, *** ***** ** ***** ******** with * *** ******-******* ****** ** came ** **** **** *** ******* ***** ****** ***’** **** ********** ******.
***** *** ** *** ********* ** submissions **** ***** ** **** ** get ******, *** *** ******* ** all *** ********** ******** ** *******. Thank *** *** **** ****!
This ***** ** *** ************ ****** ** ***:
**** ****** ** *******; ************ ****, cats, ***...***** (****** *****!)
****** **** ****** *** ****** ****** (for *** **** *** ********... *** as ******* ** ** ******.)
****** **** **** *******
****** ** ********
**** ************* **** **** **** ***** time. **** *** ** * ***** setback, *** **'* **** *********. *** cult ********* ****, *** *** **** part, *** **** **** ************* ** a **** ***** *** **** ****** be *******.