I’d like to provide an update to this thread. After reaching out to our VEX Competition Support Manager (for Canada), I confirmed that:
Worlds registration will be $975, and is not included in the $2500 season registration fee
It will not be possible to register without paying the full price for the full system. The $2500 option is the only one.
Based on this information, I think we can have an informed conversation about why registration for the pilot season may or may not be worth it for various team organizations.
I would really like to ask as to why the VEXAI electronics come with registration. This seems very inefficient as:
If electronics break, you can’t get a replacement unless they are sold on the website
Next season if you register, you will get a bunch of additional unnecessary electronics that other teams may probably want/need. As an addition, you’ll end up paying more money just for electronics you already have.
It inflates the appearance of registration. High registration cost makes VEXAI be a turnoff for many teams, which would likely lower the amount of people registering for VEXAI (Although we have Covid-19 effecting registration, I think making registration be pricy would make it even worse)
With this question answered, I think my first question becomes all the more relevant:
With the confirmation that worlds registration is not included, about $1000 of the $2500 registration fee is seemingly unaccounted for, and gives the appearance of a $1000 markup on (mostly) off the shelf equipment. @Jim_Crane@DanMantz I think an official break down of what the registration fee is paying for (software R&D?) would go a long way towards combating this appearance and would do a lot to help teams understand this price of entry into VAIC.
The $2500 VAIC registration fee includes the hardware as discussed previously in this thread. It also includes the software (including software development costs) and registration fees for the REC Foundation to support this new VAIC competition.
The effort to develop VAIC and the infrastructure to support it is significant. I do hope the community understands and respects this.
Can the RECF provide a team registration only option for VAIC? In addition to the current registration + suggested hardware + software bundle. It seems that given that all hardware is legal for sensing and processing that experienced teams would be much more interested in getting to make their own decisions on what hardware to buy.
It also would provide a substantial learning opportunity to students who want to learn how microprocessors and sensors work.
We all can understand that the development of a new competition will cost money, especially on that uses new sensors and software that was not previously used in past competitions. What is keeping Vex and RECF from getting sponsor ships to cover the R&D cost to lower the cost to teams? I would imagine that many companies would love to have their name associated with this competition and to help bring down the cost. It is the understanding of many that Vex and RECF have worked towards making competitive robotics affordable to all, this seems to be a step in the wrong direction when it comes to the growing field of Artificial Intelligence.
For this pilot year, there is only the $2500 registration option. Please note that we pre purchased hardware to make sure it was available for all the pilot teams. And as noted above, the registration fee includes the VAIC competition software and registration fees for the REC Foundation to support this new VAIC competition.
With all respect, we have been working hard - very hard - to find sponsors for VAIC. But this is a multi year process and many companies want to see the actual product / competition before committing to sponsoring this program. And with COVID-19, there are even more challenges to bring in sponsors. Many companies are impacted financially by this crisis and are more likely to take this year off from sponsoring any organization.
As always, thanks @DanMantz for being willing to come to the Forum and provide additional information on important topics!
While we have you on the line…
In future years, will they be setting it up to allow for highschool students to be on the same team as college students? My understanding is that at the current time teams can only be all HS or all college (not mixed). I think it would be a fantastic learning and growing opportunity to allow mixed teams, with University students able to work side-by-side hands-on with advanced HS students. A mixed team would automatically be considered a college team (just like when you have a MS student on a HS team it is still a HS team).
Also, as pointed out in the Vex AI - no hybrid U/HS teams thread, the rules appear to only mention eligibility for individual students and to not disallow hybrid teams. I, for one, was quite surprised at the apparent restriction, as I had assumed that a hybrid team would be something Vex/RECF would want to encourage rather than discourage.
Thanks for bringing up the software. Based on this and other information provided, it appears that the software is a significant portion of the apparent cost to teams. For that reason, I think it’s fair to ask a few questions about it (tagging @levipope as well):
What is the software which is being provided? How will teams be able to use it? For example, is it a fully specialized programming environment, or a collection of libraries that teams are expected to incorporate into software developed using standard tools?
How much are teams being provided with? Is it just an opaque binary, or will teams also get supplementary materials such as source code, explanatory documentation[1], and/or training datasets where applicable?
What will the licensing scheme be? If anything in question 2 is “yes”, will teams be permitted to share any improvements they make on said supplementary materials? In fact, are there any limits on sharing the software packages, and if so how will that affect large organizations?
I believe that all of these questions (and the follow-ups that I am sure they will generate) are important for prospective teams to know what they are buying into.
[1]: by “explanatory documentation”, I mean documentation that describes the how and why, not just the what; this is in contrast to “interface documentation”, such as the Doxygen website for VEXcode. Accurate interface documentation is extremely important for usability, but I believe explanatory documentation is rather important for pedagogical reasons.
As an EP / coach I was really excited to participate in coaching and hosting VAIC this year. I even figured it would be more feasible to host a small VAIC tournament than a standard one due to the covid situation. And my students spent most of the summer planning a joint HS/VEXU team. But the unexpectedly high cost (especially for HS as it requires 2 non-standard sized robots) and no joint HS/college teams means that we wont be able to participate.
I suspect there is a risk that there wont be enough teams participating this year, or sufficient competitions due to these barriers on top of the covid situation. If VAIC was set up as 1v1 at least in the first year (even at $1500 each), or HS and college teams could merge then I am pretty confident that we could participate.
These are great questions @nickmertin. I will do my best to answer and more detail will be released in the coming weeks.
1 and 2. The VAIC kit will come with an SD card that contains the Linux boot image for the NVIDIA Nano. This image contains 4 VEX applications that are the foundation of the VEX AI software set. These applications will be provided as compiled executables only and will require a VEX license key to run. These applications will launch when the Nano is started along with the web server the serves and the user interface to the VEX AI software running on the Nano. The user is not required to log into the nano for the system to work. Once the nano boots all the application will start and the data will be streamed to the V5 brain via USB serial port. No programming required, as far as the Nano is concerned. But you will also get a sample VEXcode application that collects and parses the data from the Nano so that is can be used on the V5 brain. The full source code for this sample application will be provided on a Github repository that we will make public soon. We will also provide VEXcode examples for robot 2 robot communications. These examples can be used to jump start your VAIC programs. We will provide basic documentation for both the system setup and software examples. We will not dig too deep in the explanation of the system for this first pilot year. Our goal, of course, is that this system will mature into a broad competition and educational tool that will not only be well documented but also the topic of a ton of future educational material.
All compiled software is provided under our VEX EULA. All source code will be provided under an open source license and can be shared in accordance with that license. The binary applications will require a USB license key to run so sharing the binary will serve no purpose without that key. As far as feedback goes, the answer is YES!. We want all the feedback we can get. This pilot year is all about exploring the possibilities of this system. We want to know the good and the bad. As I mentioned above we will be setting up a Github repository for our example applications. So users can use that for feedback and bug reporting along with this forum.
Our goal for this year was to send out a turnkey system that only needed software to be written for the VEX V5 brain. So teams that don’t want to dig into the Jetson Nano don’t have too. That is what will be going out with the kits. But we are not doing anything to lock down the linux image running on the nano. So we will provide the login info for the nano so you can (at your own risk, your warranty will be voided, all sales are final, blah blah blah) make modification if you would like. With that being said, our support team will support as much as we can but most likely if see that you have modified the base image you will be on your own.
This pilot year will help VEX and the REC set a path into the future of competition/educational robotics. We want teams that are committed to that purpose to be the ones that help us get there. I personally cannot wait to see how this season goes.
I realized that I probably did not go into the level of detail you were really wanting. But as soon was we are finished setting up the public Github it will contain all the information you need.
This makes sense to me from VEX/RECF’s perspective, especially for a pilot year. However, the initial feedback that I can provide, as someone affiliated with and advising multiple organizations who are interested in VAIC, is that this is not a sustainable model for future years. Teams will have a difficult time justifying the large cost of joining, especially if they are doing so for the express purpose of developing their own system, and advancing their skills by doing so. I can say with confidence that there are many prospective teams (largely, but not exclusively, at the University level) who fall into this category. This is not a reflection of the quality of the software provided by VEX (obviously I have no way to judge that), but of the overall pricing model. I think VAIC has a lot of potential as long as it’s understood that teams who are at the level necessary to be interested in it will want to have more control over what they do.
Here, I see a few potential issues cropping up. Firstly, given that teams have no way of judging the quality of the software before dropping US$2500 on the registration, and by the sounds of it have no way to improve it (it’s either use the VEX software or start from scratch and void the warranty), I personally could not justify buying in. I understand why it makes sense for you from a business point of view, but given the teams’ perspectives (including their experience with the Vision Sensor), I can assure you that many of the teams looking into VAIC right now were going with the assumption that the VEX software wouldn’t work well enough out of the box. Requiring them to effectively pay US$1000 for that software—and voiding their warranty if they try to build their own system—is not going to help bring them on board. Moreover, in my view this is counter-productive to the educational goal of the program, as it eliminates students’ abilities to learn by tinkering and building on existing systems.
Secondly, there’s the physical license key. I’m not sure exactly how this is going to work, but given the widely documented false-positive failures of DRM systems in the wild, this is the biggest thing that jumped out at myself and others as questionable. What recourse do teams have if suddenly, this several-thousand-dollar system that they were required to purchase in order to compete stops working on the morning of competition. What if the physical license key breaks, or is lost? Can teams get a new one? Obviously they don’t want to (and shouldn’t) have to pay exorbitant amounts of money to do so, but how else is VEX going to protect the licensing scheme? This is a much larger hurdle (with business strategy playing a disproportionately large role) than teams just buying a backup VEXnet key.
I want to be clear that I’m not trying to shoot down all the effort that has been put into VAIC. I want it to succeed as much as anyone at RECF and VEX does—especially as a potential competitor myself. Rather, I’m trying to get all the issues that I can see off the bat out in the open, to prevent them from holding back the program’s potential. Right now, it seems to check all the boxes as far as RECF and VEX’s business requirements, but does not check a lot of the boxes for the community to invest in it, which is just as, if not more, important.
This at the University level makes absolutely no sense the NANO code should be open source and likely is based already on Opensource tensorflow, OpenCV etc if the goal is to get University students to get attracted to VEX then opensourse is a must and will as such create a community of followers and advocates, otherwise do DuckieBot, Amazon Race Car and on and on which all have great competitions and the entry price is less then $300 - I truly think VEX is missing an opportunity here.
At this price point, and given the closed source nature of what’s being done, I really do struggle to see what appeal this program will have.
To field a team from scratch, you will need
Registration at $2500
2x V5 competition starter kits at $1k each, for $2k
A field perimeter at $800
A field element set at $500
Field tiles at $200
For a total of $6,000. This doesn’t include any event registration, and it also likely isn’t enough to field a good team with just the starter kit.
For context, I could buy a used electric golf cart, a LIDAR, a computer, etc and start an autonomous vehicle club/program thing for around the same price. Or as was pointed out, I could compete in the Amazon race thing with a kit that costs $400.
This is to say nothing of the fact that all the interesting parts dealing with machine vision are locked in a black box that you’re not allowed to touch, and which are hard coded (presumably) to only work in the context of VRC. I couldn’t use this same kit for some other cool side project, it literally only works in the context of the VRC game, I can’t look at or modify the machine learning/vision stuff to learn from it, it’s just a black box that says “this is where you are, this is where the game objects are.” Maybe you have solid business/legal reasons for keeping this closed source, but it just seems to me that doing it this way removes a lot of the chances for any actual learning.
Certainly in the context of a job interview, if somebody interviewing told me “oh, I worked on this machine vision thing” I’d find that really cool and want to hear more, but if it turns out that they just used a black box they had no understanding of, I’d probably move on to the next candidate.
Considering the bots are the same as U-bots, I figure RECF is thinking mostly established U-teams would be doing this, so it’s just the $2500 for the majority of participants.
Yeah, that’s fair. Although (and I could be wrong) I don’t think all that many people are doing VEX U anyway, so they’re fracturing an already small competition into an even smaller one? I guess field the same bot in both competitions if you’re really excited about both, but then you can only go to one at worlds.
Also, long term even if they got it down to a more manageable, say, $1500 to start that’s still $5k initial buy in. I could see it being worth it to some people if you could actually do something with the cool machine learning stuff they have going with this system, but if it’s just a black box… I dunno, I had a lot of cool opportunities in college, even ignoring cost I don’t know I would have found this worth my limited free time.
Normal team registration is 150$. The hardware provided is worth about 1200$. (after the people, like Nvidia/intel who actually spent years designing it get their cut)
A 500$ registration only option would still be 350$ extra for “development costs” and the teams don’t even need to be provided software for their trouble.
A 1000$ registration + software but no hardware option would be 850$ towards the “development costs” and offset the sunk cost of RECF buying hardware.
Everyone here is okay with supporting VEX/RECFs efforts to create this program. But would rather not spend 1200$ on paperweights they wont use. I am trying to meet you halfway and I would appreciate it if you did the same.
