Wow, I interviewed with Starsky for a summer internship position back around the beginning of December. The interview went well, but I never heard back from them. I just assumed that I didn't quite clear the bar, but it looks like they got the news that funding wasn't coming (timetable from the article matches up) and just shut hiring down.
During the interview I asked my interviewer about Starsky's long-term prospects, and he said that they would have near doubled in size come summer. Lol
> “Our approach is taking the drivers out of the truck and have them working remotely to make high-level decisions while the truck is on the highway,” Seltz-Axmacher told FreightWaves in May 2019. “To be honest, I don’t think that a super-computer can be built that is smarter than a truck driver.”
Probably why they didn't get funding. VCs don't want to hear that (even if it may be closer to the truth than anyone wants to hear).
This seems like a setup that while a logical architecture decision, would be reliant upon really rock solid mobile broadband on the routes driven by its trucks. Like, seven nines reliable. I saw videos of their test setup with a driver at a desk workstation/"cockpit" which was driving a remote truck.
Imagine if a human needed to make a decision with less than a few seconds of reaction time, at freeway speeds, and the LTE network where the truck happened to be at that moment was having a hiccup.
It's much worse than a cargo drone that might have a 12-14 satellite GPS+GLONASS lock, that could be programmed to stop and hover in place, or retrace its flight path breadcrumb style back to base in the event of brief loss of WAN network connectivity. It's something with possibly 20,000 kilograms of mass moving at 70 mph with live humans in traffic all around it, and a lot of inertia.
It's an external technical problem that the company itself can't solve, because they're not in control of LTE/5G network operators, they don't make LTE radios/modems/antennas, etc. The only places that I'm aware of using fully remote driven trucks are giant open pit mines in western Australia, where the mining company fully owns and operates in-house the network linking the trucks.
I personally know very little about self driving technology, but working in network engineering, the idea of relying upon the normal publicly-accessible nationwide LTE network operators for something that could be a life-critical safety function really scares me. Even with a max-priority service like Firstnet on ATT it would be sketchy.
The trucks could have AI that worked most of the time. Not well enough to rely on for normal operations, but good enough for a fallback in case of losing connectivity.
Some made-up numbers to explain my point: Consider a situation where the remote operator will make the right response 99.9% of the time, failing 0.1%, and the AI is 20 times worse, making an error 2% of the time. The actual error rate will be (0.1% * connection reliability) + (2% * (1 - connection reliability)). For 3-nines reliability, which is pretty poor, the error rate will be about 0.102%. For 4-nines it will be 0.1002%.
Those sorts of figures would effectively pay for themselves in terms of safety, once you factor in the benefits for monitoring and work conditions for a desk-operator vs a physically present driver, and they don't require either a particularly reliable connection or a particularly good backup AI.
The exact cases where human reaction times fail are already taken care of by local subsumptive AI† in cars. For example, collision avoidance is “AI”; lane-keeping and adaptive cruise-control are “AI”; etc. For that matter, an automatic transmission is “AI” in the sense of abstracting away a clutch and gears under a prediction of when you’d want them used.
† ...which doesn’t actually imply anything to do with Machine Learning. Your computer downclocking when it overheats rather than catching on fire, is an example of “subsumptive-architecture artificial intelligence”, even though it’s a hand-coded algorithm doing the subsumption. The “artificial intelligence” in this case refers to the emergent property of the layered subsumptive processes, not to the processes themselves.
I think their goal was to have highway driving completely handled by the self driving systems with maybe call outs if it gets into a situation it can't figure out (weird construction zone or accident it needs to move around). For covering the gaps existing technologies might even be enough, a combination of auto breaking, adaptive cruise control and lane keeping would be a pretty solid gap fill.
There's several things that could be done in the city if it experiences drop outs. A creep mode where it moves quite slowly till the network improves would let the lower bandwidth or latency connection still be used safely.
Personally I always thought the answer would be self driving on highways to cargo hubs for delivery by human drivers. Or maybe self driving trucks with sleepers so if it got into a confusing situation it could stop/pull over and wake the driver to deal with it and allow the driver to sleep while the truck keeps moving.
>maybe call outs if it gets into a situation it can't figure out
How would it predict that these situations are coming with enough time to hand off to a human driver? I think it would need some form of crowd sourced or externally gathered data. The truck itself couldn't gather this in time.
>A creep mode where it moves quite slowly till the network improves would let the lower bandwidth or latency connection still be used safely.
Sounds like a traffic nightmare.
>Personally I always thought the answer would be self driving on highways to cargo hubs for delivery by human drivers
I do agree with this. I firmly believe that full self driving will probably never be possible for anything except for Interstate/main highways that are well maintained and very standardized in their construction, markings, etc.
There are source out there for accidents, traffic information and construction where the system could pretty reliably predict when it's going to encounter an odd traffic pattern that would be better handled by a human than the automated systems. In cases where there's a decision needed extremely quickly we're generally talking about an accident or near accident that relatively simple existing assistance systems like autobraking and obstacle avoidance would handle.
Trucks are already really slow compared to traffic inside cities and in the rare cases where they're in this mode on a highway they're probably already in a slow traffic area because they're dealing with construction or accident areas.
Well there is good reason they don't want to hear it. It's a stupid idea. The last thing you want is some driver 300 miles away on a computer not paying attention, or with the ability to crash a semitruck, or with the uncertainty that your entire fleet goes down if your network goes down, or if someone finds a weakness in your security and starts crashing all your trucks. I mean whatever moron was responsible for this path in this company (likely the CEO) should be fired and replaced. It looks like at this point they will just end up liquidating this company anyway.
But to be honest they haven't done anything too notable within the last year (Plus.ai is pretty much the biggest thing I can think of right now and maybe Ike in the trucking space).
Really interesting to see how everything in the AV space plays out: Deepscale/Drive.ai were both aquihires, Cruise/Torc/Argo/Aurora were bought out mostly, and the others are either just starting (Ghost and Voyage) or doing something that aids them (Renovo).
So essentially: either you're bought out by an OEM or funded by VCs hoping to get acquired (I don't think you can surivive in this space without being bought out or aquihired).
> But to be honest they haven't done anything too notable within the last year
Per the article:
> The company was the first to complete a fully unmanned truck trip on a 9.4-mile stretch of public highway and claims to have set a speed record for an unmanned truck, reaching 55 mph, in June 2019.
I'll admit that their telepresence human driver isn't ideal for self-driving vehicles. Its strategic merits, and product merits are very arguable. But it does follow a familiar model in automation progression, and I'd imagine that quite a few investors would have been interested. Likely just came down to timing and potentially valuation or other details. Can really hurt when an investor pulls back at the last minute.
Zoox was supposed to announce a new funding round by the end of last year, and still crickets. The further into 2020 we get the more I think they might not get any more money.
I actually loved what they were (and still are doing).The only thing that concerns me about them is this:
1. Ousted one of their CEOs the day they got funding (something dirty like this)
2. Making their own car (which I guess isn't too far fetched, but Cruise seems to be much further along even though they did the whole thing in secrecy)
Just checked out the demos on their website. I have always been curious what graphics they use to represent what car sees. Are there any open source versions of these out there to explore ?
Stiching cams like this has started to hit the back up camera market. Takes a few hours of setup. Stop at an RV dealer and check out the cams systems on the $$$$ units.
Interesting, it looks like I am out of the loop, haven’t heard about most of these except cruise and aurora. Where do you think Zoox falls on this spectrum?
Website still says they're hiring. They signal robust growth, you pay for it in wasted time. I wonder how many jobs I've applied for have been like that.
I know of deep fake employment opportunities designed exclusively for identifying opportunities for IP larceny. This is a huge reason for locking down everything if you have any relatives of important people in your organisation as interns. This is the oldest privilege escalation hack I know of, and you might want to see if your outside legal will support a fake interview process for your own people via a headhunter. Reciprocal and wholly legitimate arrangements exist like this in many industries. It’s frightening how much we found we were leaking over family breakfasts in seemingly innocent ways like revealing meeting details and simply mood responses to eager-junior inquiries. Send your sons and daughters to start their own business and learn to value poker playing before they get inside the “family” business. BillG played poker rather well I’ve read...
The fake vacancies I had in mind are not so malicious. But for the candidates, in aggregate, they are a waste of precious time and energy.
- Company has still opportunities on their website, in order to look healthy, when they are in fact laying everybody off (as is the case in the OP)
- Company publishes an offer for legal reasons, but in fact wants to hire a cheaper resource from abroad once they've "proven" that they "couldn't get any local talent"
- Company publishes non-urgent offer, in the hope to fish for resumes, and/or waits for years for the "perfect" talent to knock on their door
- Company forgot to take down vacancy, someone got hired 3 months ago
Those are all real problems but there is an insidious one where you pretend to poach an employee from a competing firm but instead just do everything you can to brain dump them during the interview.
I wonder how much of that information would be actionable though... I remember a story I heard at my last employer, one of our younger designers had brought some work home and his father, a VP for a rival company, saw his screen and exclaimed "This makes so much more sense! Why aren't we doing this?!"
But of course, they didn't start doing it like that because they had a huge workforce already trained on their own process and the juice wasn't worth the squeeze. I'm not sure how I would actually exploit a rival's meeting minutes even if I had them and no scruples.
Got a buddy in HR, he mentioned he regularly floats fake listings to see:
1) what kind of hits they get (e.g. lots of over- or under-qualified)
2) what the salary requirements are relative to the hits they get
3) how effective are the ads in terms of writing, conveying what is on the table without saying too much, etc.
4) to make the company look busy / growing / large
5) to stay ahead of the churn, since they're large enough that they may need bodies in a few months (e.g. a poor performer is on the way out / gettin the axe, or a contract is up in 3 months) so they can get a couple of prospects on the hook and do a slow negotiation until they're ready to hire, in which case they make an offer at or near market rates
What it probably means in this case is that they had a term sheet, but that during due-diligence (1-3 months between term sheet and signed contracts) they pulled out. It's possible they found things they didn't like in DD. It's not uncommon to find oddities, but normally not a big issue at all.
It's a really bad move. It looks very bad to companies who might be seeking investment, and could get a VC a name for causing issues in deals, so they only do it when they really need to. It typically means the company closes down as most companies raise when they have ~months of runway left, and raising a new round would take too long in most cases.
Why couldn’t they raise more money or get acquired? I thought self-driving car startups were worth $1mil per engineer (at least!) in an acquihire let alone semi-functional tech. Was it a scam?
> “We beat Waymo. We beat all the big OEMs. We’ve beat just about the whole industry,” says Stefan Seltz-Axmacher, founder and CEO of the San Francisco based start-up.
I find this statement pretty disingenuous when Waymo is trying to tackle the part of the problem Starsky isn't able to solve with their own self-driving tech - driving without a human in the loop for the last mile.
You should check your figures. I believe Drive.ai, one of the leaders in the industry, was acquired for about $500,000 an engineer last summer. The numbers you are thinking of are more appropriate during 2016-2018 when the hype train was at full speed. Last summer there was a palpable feeling that the industry was beginning to 'consolidate'. I would not be surprised if all of the non-OEM, non-Google backed AV companies go bust soon.
During the interview I asked my interviewer about Starsky's long-term prospects, and he said that they would have near doubled in size come summer. Lol