TruckingInfo reports Loadsmart and Starsky Robotics Make First-Ever Digital Freight Delivery Via Autonomous Truck
According to the two companies, the integration of Loadsmart’s AI-powered pricing and load matching technology with Starsky’s API meant no human intervention was required. The historic initiative is part of a larger strategic partnership which paves the way for the future of trucking: autonomous brokerages dispatching freight to autonomous trucks without human involvement.
Loadsmart said it was able to connect its network of customers with Starsky’s fleet of regular and self-driving trucks by integrating Loadsmart’s Automated Dispatch API with Starsky’s Hutch API.
As a result of the partnership, Starsky is able to dispatch its trucks automatically without human intervention, while Loadsmart can expand its ability to automate the shipping process from quoting to booking to delivery to help its clients move more with less.
“For the first time ever, the advances that seem obvious for the ride-sharing services are coming to trucking,” said Stefan Seltz-Axmacher, CEO and founder of Starsky Robotics. “It’s not uncommon for a traditional trucking company to have five full-time employees involved in dispatching each truck for each load. By integrating e-brokers like Loadsmart, we are eliminating all back office human intervention and making the shipment process seamless, while focusing on ensuring the safety of driverless trucks. With Starsky’s Hutch API, which was also announced today, we will be able to autonomously dispatch autonomous loads on a regular basis.”
Delivery Details
Trucks.Com has additional details on the delivery.
Digital freight broker Loadsmart and self-driving truck developer Starsky Robotics completed what they say is the first autonomous dispatch and delivery of freight. The team paired in late July to book and deliver a load of corn in Texas with minimal human involvement.
Loadsmart digitally priced, tendered and booked the shipment. A Starsky self-driving truck picked up and delivered the raw corn to a customer in Grand Prairie, Texas.
Such partnerships are likely to expand as freight and logistics becomes increasingly digital, said Cathy Morrow Roberson of Atlanta-based Logistics Trends & Insights.
“Eventually I see the marriage between digital freight brokerages and autonomous trucks,” Roberson said. “It makes sense from an efficiency and timing perspective. And it could be beneficial as the trucking market continues to struggle with attracting and retaining drivers.”
Mass Adoption
There are two things holding up mass adoption.
- National Regulation
- More Testing
Requirement two is proceeding nicely.
Starsky plans to begin commercialization next month. To find potential problems an autonomous fleet might face, the company has a 6-person team operating a 40-truck fleet.
“We want to triple the size of that fleet by the end of the year, but we don’t want to triple the size of our operations team,” Seltz-Axmacher said.
He plans to build 25 autonomous trucks by the end of this year. To limit human involvement, Starsky intends to use its proprietary application programming interface, or API, to dispatch the trucks for tests.
“What we are doing is creating an API that brokers can use to negotiate for additional capacity and then hire that capacity without anybody talking to anyone else,” Seltz-Axmacher said.
Intense Competition
Competition in this space is intense. I have lost track of the number of competing in this space. It includes all the major car manufacturers, Waymo (Google), Otto, Loadsmart, Starsky, Amazon, Uber, and Lyft.
That’s what guarantees success, sooner rather that later, despite the poor current technology of Uber.
Robertson gets it correct with her assessment “Autonomous trucks will become more the norm than the exception.”
Inner-city truck deliveries will still require drivers, for a period of time, but rural deliveries like the one above and hub-to-hub interstate deliveries by autonomous truck will soon be standard.
100 Pigs Die
On July 19, an Overturned Semitruck Closed a Freeway Ramp in Louisville. On July 17, More than 100 Pigs Die After Truck Overturns on I-65 Ramp at Spaghetti Junction near Louisville.
Benefits
- Accident rates will plunge. Nearly all truck accidents are driver error.
- Insurance rates will drop.
- The cost of the driver will be eliminated. No more union dues.
- Trucks can drive 24-7. They do not have to sleep.
The benefits are overwhelming. Tremendous benefits coupled with intense competition ensures success. Within a year or two of national regulator approval, autonomous semi-truck deliveries will be the norm.
Mike “Mish” Shedlock
The promise and hype of this tech is nothing but the failing alchemical desperation of mankind trapped by his own stink. There isn’t a single car in production now that has over a 79% reliability rating. NOT ONE! And 79% reliability is just for a hand full of high end very expensive cars. If your car breaks down now no one is killed or maimed. Do you think humans can produce self-driving vehicles with enough reliability to avoid a 20% fail rate? Think again bucko!
The trick, as in all of engineering, is to cordon off a much more limited environment than the fully general one of the world as presented raw, and then build a solution for this environment.
Interstate-hub to interstate-hub is one such cordoning off. Separate lanes/roads altogether would be even better, but even a shared interstate may be limited and predictable enough to let machines compete with human operators for efficiency.
Auto-driving trucks going truck hub-to-truck-hub are the most natural fit for implementation and will happen first. All other applications will be much more difficult and may be decades off until (when and if) autonomous autos dominate traffic. The 24 hour operation is the deciding factor, even if the driving needs to be done at lower-speed restricted to the right lane. Weather and unexpected events will also not overcome the advantage of a 24 hour-a-day operation.
Things that need to be worked out: refueling of auto-driving trucks, relocation of some hubs out of metro traffic areas, out-thinking mischief of others.
As a long time reader of Mish I remember when Mike declared all truck drivers would be out of work within three years because some company I’ve never heard of promoted a souped up prototype resembling a golf cart to deliver goods “the last mile”. That was more than three years ago.
I do appreciate Mish adopting the logic I’ve argued countless times in the past of a hub-to-hub automated interstate method as his own even if he was pulled to it kicking and screaming.
Let’s examine his benefits of automation:
* “Nearly all truck accidents are driver error.”
True, but not like you think. Most truck accidents (80%) are caused by CAR DRIVER error. That will not change with automated semis.
link to ccjdigital.com
If you’re going to lose your insurance, your career, and all means of supporting yourself, you have skin-in-the-game. You drive carefully. Semi drivers drive 10 hours a day, every day. Car drivers only occasionally drive 2-3 hours/day.
* “Accident rates will plunge.”
Negative. There will be an initial spike in accident rates as semis are introduced resulting from the unforeseen side effects and unintended consequences of automation, but the mirage of perceived savings will keep drawing money and tech like moths to a flame.
* “Insurance rates will drop.”
There will be some productivity gains with long haul automation but you’re kidding yourself if you think an unproven technology at such a dangerous intersection with humans won’t have massive litigation expenses.
* “The cost of the driver will be eliminated. No more union dues.”
Only 2% of truck drivers are unionized.
Those “Over-the-Road” (OTR) long haul drivers will *slowly* be eliminated but they will instead take jobs closer to home as local “last mile” drivers. This will actually work in their favor as they’ll get more home time. Drivers, and their expense, will most certainly not be eliminated in the next 15-20 years because that “last mile” problem has no easy answers.
* “Trucks can drive 24-7. They do not have to sleep.”
No, but they still need maintenance. Excepting the engine, most semis don’t have sensors attached to every critical piece of the truck and trailer. Who is going to pretrip these automated trucks? If a hub seal is broken and runs dry causing a tire fire, how does the truck know? If the truck doesn’t know, that fire will spread to the adjoining tires and then to the trailer itself. That should be fun at 65 M.P.H. .
I can see a new business opportunity. On call truck drivers and repair. For the times a truck breaks down or gets stuck in a tight spot. They call a local company to get the truck back up and running. There will likely be something like a code that will allow a driver and/or repairman to get access to the truck.
When remote key-fobs first came out, it took car thieves less than a week to get copies of the universal key fobs that car dealerships had.
If there is a special access code for repairmen or emergency drivers, thieves will have a copy faster than you can type a reply to this message.
Its always easy for armchair quarterbacks to say how simple it is to do such and such. In the real world, its more difficult and takes a lot longer than the experts predict.
Now excuse me, I have to get in my George Jetson space car and fly off to work.
Or, since we’re in living in Hypeistan and all, just dispatch a self driving truck fixing robot to do the job…. Diagnosing and fixing a broken truck, is an infinitely less complex problem than operating one in in a truly general traffic environment, after all….
Anybody have any ideas on how to turn 3-4 million ex-truck drivers into consumers?
Yes, let them take the bar exam, become trial lawyers and help sue the truck companies whenever there is an accident.
Soulless, faceless robot vs crying human family that just lost a loved one — in front of a human jury.
We don’t even need to look at the evidence to know what the jury will decide in most every case
And how is that different now. Just watch afternoon TV – all the ads for personal injury lawyers. That is what insurance is for and is a cost of shipping…
Truck drivers tend to win most lawsuits.
Trial lawyers get cost (including their own salary) or a percent of damages. Lawyers don’t have to win to profit.
Auto-bots will lose more cases, because a human jury will sympathize with a human plantiff a lot more than a soulless computer.
Too many people, especially the ones who end up on juries, think computers are infallible. And they think, like you seem to, that insurance payouts are free money with no costs.
Higher shipping costs to pay for endless lawsuits is a good business reason to stick with human drivers — even if the tech were able to live up to the hype
Almost all personal injury cases are done on pct of award. Typically 25-33%. Depending on what the state allows. And there is rarely a jury, It’s settled between the lawyer and insurance company. Neither side has an interest in going to court.
RTFD — I already wrote ” or a percent of damages” and I included the fact that most lawyers get their costs reimbursed if they don’t win. Working completely contingency rarely happens in reality. Lawyers will advertise the contingency possibilty, but they generall wont take those cases unless they believe the client is likely to win. Lawyers do not work 100% pro-bono (that means for free); they work with the expectation of winning a big settlement.
Further, you missed the point completely. No insurance company wants to insure a soulless machine that no jury can relate to.
The added business cost of a losing trial (or having to settle before trial) is reason to stick with human truck drivers — at least for the next couple decades.
First, they will need to get the software working better. After that, they will need to convince the public, and only after that they might be able to convince insurance companies. Its a long way off.
Human drivers are still doomed. Most accidents are handled by insurance companies and insurance companies are very good at evaluating and pricing risk. Even if the automated truck requires double the insurance cost, it will still cost far less than the human driver’s salary and benefits. Plus the equipment utilization and fuel efficiency will soar. There will always be lawsuits, but it’s not going to stop this inevitability.
3-4 million ex truck drivers will most surely have 3-4 million different ideas about how to do exactly that, should so be required.
Which is all that matters, as “turning” someone else into something, is hardly any business of anyone but those individual someones themselves.
Sure: robotics technicians.
At first, fueling operations will be human, but then yield to robots. There will be many opportunities in the manufacturing, installing and servicing of robots.
Many small rural towns will have to find revenue replacement for the dislocation of travel services and related infrastructure logistics.
All municipalities will have to grapple with the loss of traffic violation income-especially as auto-autos are adopted. (More bad news for property taxes and pension benefits.) Texas had to pass a law limiting abusive towns to not more than 50% of their budget could be raised by traffic fines.
“That will take a decade or two and probably investments of $300-500B, not a trivial amount.”
It will take 3-4 years – savings is too great for commercial enterprises
What may take a long time is non-commercial
Perhaps 2028-2030
Call it 1 decade but perhaps that is far too pessimistic
Especially in the wake of the $15/hr minimum wage fiasco, which is closing restaurants and small businesses nationwide, there is absolutely zero hesitation on the part of any business to substitute capital for labor — to replace McDonalds burger flippers with robots or replace truck drivers with “auto-bots”
The legal liability issue is the problem. There are too many lawyers to ignore what happens when a jury looks at a soulless robot with no face, and then looks at the crying family of someone killed in an accident involving an auto-bot. Even if the auto-bot wasn’t at fault, it is a soulless machine and here is a crying family that just lost a loved one.
The law isn’t about justice or the facts. If facts had anything to do with the law, Comey and Brennan would already have orange jumpsuits
OTR rigs are replaced everyday. Few fleets have old rigs operating beyond their warranty period now. The switchover will occur organically and seamlessly.
“I wish to see how they prevent hacking into such trucks to avoid someone making them driving into crowded place. “
Yea – right easy as pie
Iran hacked into the guidance system of a military drone, which presumably had better cyber security than a trucking company ever would.
And an easy form of “hacking” was invented in the Netherlands hundreds of years ago – sabotage. What stops unemployed truck drivers from slashing tires? (plenty of other ways to mess up a truck besides tires)
I envision problems – including one Mish did not mention. Union push-back – including sabotage. But that will only slow things a bit – not stop the revolution.
Truck drivers vote, auto-bots do not. The teamsters have bought off members of Congress for many decades. What stops them from passing onerous certification laws against auto-bots at the national level?
As I already mentioned, slashing auto-bot tires is a sure thing. Yanking air brake hoses. Stabbing a screw driver into the gas tank to let all the fuel leak (and requiring a human repair man). Super glue the lock / door on the back of the rig so it can’t be unloaded at destination. Spray paint the cameras the auto-bot uses to see where it is crashing.
And bribe members of Congress to make sure teamster jobs are protected.
The technology isn’t fool proof — it will take another 5-10 years to get auto-bot technology able to handle expected road circumstances. None of the tech companies (Google or Tesla or anyone else) are located in snow prone areas. How many Google programmers know how to drive in snow?
Once the tech gets worked out… It will take many many years to work through the human problems of auto-bots. Sabotage, hacking, and the big one: trial lawyers
Yeah right – let’s compare Tesla to Waymo
Waymo has the backing of a cash cow (google) – and they are taking things very slow.
Tesla needs Elon Musk to find enough suckers to “invest” before the SEC prohibits Musk’s companies from violating securities law any further
If the SEC allows Musk to ignore securities law, everyone else will too. For a federal bureaucrat, being irrelevant is a nightmare they cannot allow to happen.
The only comments are negative because there’s no need to advocate for autonomous cars anymore. They are as inevitable as the weather.
The promise and hype of this tech is nothing but the failing alchemical desperation of mankind trapped by his own stink. There isn’t a single car in production now that has over a 79% reliability rating. NOT ONE! And 79% reliability is just for a hand full of high end very expensive cars. If your car breaks down now no one is killed or maimed. Do you think humans can produce self-driving vehicles with enough reliability to avoid a 20% fail rate? Think again bucko!
And HAL 9000.
AI is always inevitable, to those who are themselves too clueless to even attempt being the ones building any of it.
Just like that guy named “we,” can always easily solve all political and economic problems……
Thankfully computers are always secure, and can’t be hacked into. Otherwise, image what might happen.
Yet another Tesla auto-driving crash in the news. The other players, including Google, have drastically scaled back their aspirations. National auto-driving might happen some day, but its still many many years away.
And Mish somehow forgot to mention a HUGE road block to auto driving vehicles. In a society that is absolutely infested with trial lawyers, it is naive to think someone won’t be held liable in a court of law when these things crash.
No blog post that ignores this legal liability has credibility, and Mish isn’t a lawyer
I wish to see how they prevent hacking into such trucks to avoid someone making them driving into crowded place. Huge competition and short supply of skilled programmers guarantees such possibilities being possible.
If it becomes a problem, there will be a solution. Maybe a guy in Las Vegas in charge of 50 trucks – monitoring and notifying the appropriate folks, if something goes wrong.
“Sir, the Iranians appear to have taken control of our multi-million dollar drone and are piloting it to land intact at an Iranian air force base where they can reverse engineer everything.”
I’m sure “some guy in Vegas” is going to have much higher level of cyber security than a military drone. Or maybe a minimum wage dispatcher will save the day… Let’s assume the dispatcher doesn’t have a Vegas phone number that forwards to Mumbai.
“If it becomes a problem, there will be a solution.”
And if it becomes one billion problems, there will be one billion solutions….. Ditto emergent problems arising from interactions between solutions to some problems, and other problems.
That’s what open ended complexity looks like: Lots and lots and lots of small (or big) problems which each require their own testing-fixing-testing-fixing…… cycle.
Or, going by the “limitations” of the only scalable metasolution to such problems demonstrated so far, biological evolution: Testing-fixing-SETTLING…. may be more appropriate. As there is precious little precedence for general solutions to truly open ended problems, which does not involve much greater tolerance for random failures/accidents than what more traditional engineering disciplines are accustomed to. The individual-actor flexibility required to make a “solution to the traffic problem” truly general, may very well be great enough to simultaneously demand the kind of tolerance for individual “driver errors” which the current crop of AI hypesters, and their uncritical groupies, like to pretend is some form of easily engineered out, purely accidental orthogonality.
I think we will also see many self driving trucks driving at night. A human driver will drive the last few miles. This will mean better utilisation of the road infrastructure.
Another benefit, double edged sword, fewer trucks needed.
This will feed back into the automotive-industrial complex in a deflationary way. Fewer factories etc. Demand reduction across the spectrum.
So – fewer workers, less end demand, less industrial demand.
The potential deflationary impact of technology will grind away for a long while yet.
Providing that it actually does work, the benefits will be overwhelming. However, it’s not possible to transform an industry with 3 or 4 million trucks in a year or two. That will take a decade or two and probably investments of $300-500B, not a trivial amount. Most will be content to sit back and wait for a while to see how it works out.
Good point and I agree. However, I don’t expect that period to be that long. Maybe 2 years, maybe 3. These trucking companies will ultimately offer insane incentives to get the big shops to consider the change, knowing there is an up front cost and progressive savings over time, ie, loss of human means rising salary, riding health benefits, etc etc etc over time that is no longer an expense item. That’s pretty compelling.
Just recently I’ve been consciously observing labour in places I go to – all the usual – shopping, tolls and lots of different activities (weekly necessities). The coming avalanche of labour replacement is scary.
Combined with older people remaining in the workforce means kids need to get up the skill ladder if they hope to get a start.
More competitive, basic opportunities being closed down.
It’s bound to show up in the politics.
In other news, a Tesla in Moscow, apparently on autopilot, crashed into a tow truck, bursting into flames.