Is this 9 trillion industry about to be disrupted by AI?

*Honk* 

*Honk*

It's time for another mind-expanding newsletter about the technology on everyone’s lips - artificial intelligence.

Today I am exploring how truckers and their 18-wheelers are about to be disrupted by a bunch of nerds out of Silicon Valley.

Today’s headlines:

• AI in the trucking industry

  • Background

  • The state of the self-driving industry

  • Self-driving vehicle classification

  • Applications of AI in trucking

  • Implications of AI in trucking

• AI art of the week: The future of transportation

• DuckDuckGo jumps on the generative AI craze

• AI experiment of the week - ChatGPT helps Data Scientists

• Quick bites - interesting AI news

Put on your truckers cap, and let’s go…

Background

According to Statista, the global logistics industry in 2021 was worth over $8.4 trillion and is expected to grow to $14 trillion by 2027.

The logistics industry encompasses cargo transportation services by land, air, and water.

Global logistics costs represent about 10.7 percent of the global Gross Domestic Product (GDP).

In America, there are over 900,000 employed truck drivers and about as many self-employed.

In Europe, there are estimated to be over 2 million truck drivers.

In America, as much as 72.5% of the freight is moved by trucks, according to Zippia.

In the EU, that number is even higher, at 77%, according to Acea.

According to Zippia, groceries stores in America would go out of stock in three days if trucks stopped driving.

I am sure we would have a similar situation in the EU.

Logistics and truckers are like the plumbing of our economy.

They are essential for things to go around…

The state of the self-driving industry

More and more companies are entering the self-driving arena, trying to develop technology that will become the preferred solution for self-driving trucks.

And it is not only the traditional truck companies like Daimler Chrysler or Volvo working on self-driving solutions.

Tech startups are emerging, developing brand-agnostic solutions.

At the same time, governments are working on laws and regulations for letting self-driving trucks on the road, including safety standards, etc.

Many in the logistics industry are excited about the technology with its promises of improved productivity, fuel efficiency, reduced costs, and better traffic management.

But truckers are worried about what it means for their jobs.

Understanding the levels of self-driving vehicles

Self-driving vehicle classification

• Level 0: No automation (older cars, typically)

• Level 1: Driver assistance (like adaptive cruise control)

• Level 2: Partial automation (two or more advanced driver assistance systems, like adaptive cruise control and automatic emergency braking. Tesla, despite popular belief, is still at level 2).

• Level 3: Conditional automation (capable of taking full control and operating during select parts of a journey - no cars as Q1 2023 is at Level 3).

• Level 4: High automation (capable of completing an entire journey without driver intervention, even operating without a driver at all)

• Level 5: Full automation (complete hands-off, driverless operation under all circumstances. No steering wheel or pedals needed)

The trucks and technology we discuss here are on levels 2 to 4.

Applications of artificial intelligence in trucking

I decided to write about AI in the trucking industry because I came across an article about Embark Trucks, founded by two friends at University in 2016.

They rose to “unicorn fame” and went public via a SPAC merger in 2021 at a valuation of $5,1 billion.

Doing research for this article, I learned that Embark almost went under in 2022, the share price has dropped 98%, they fired 70-80% of its staff, and is trading below cash balance.

They are still alive and working on their self-driving tech, albeit delayed due to their financial situations and the venture markets “drying up.”

Embark Trucks is developing platform-agnostic self-driving technology that can be integrated into any truck. That means both the AI software and the actual hardware that goes on the trucks.

They utilize advanced cameras, radars, lidars, GPS systems, AI, and existing map solutions to make trucks self-driving.

The AI is able to read the environment around the truck 360 degrees, with millions of data points analyzed a second, enabling it to make accurate and safe decisions while navigating the roads.

Embark Trucks have identified the long haul part of trucking as the problem to solve and trying to develop a “transfer hub model” for self-driving trucks.

In short, they want to solve the problem of shortages of truckers and safety on the roads (road fatalities) by developing self-driving trucks that haul goods between large logistics hubs.

The first and last-mile deliveries, basically the transport from production to logistics centers and from logistics centers to customers (short-haul), will still be done by humans in the eyes of Embark.

Their solution is a Level 4 solution meant to make trucks 100% self-driven without a driver in the cab or any intervention - something they hope to reach in 2024.

Embark Trucks was the first company to start testing self-driving trucks on public roads. They have performed many successful drives and continue to develop and improve their AI technology and hardware solutions.

Because of their rise to fame, there is quite a lot of information about them and their solutions.

The video below is well worth a watch! It gives you insight into the application of AI in trucking and the challenges they face.

In my opinion, it makes sense that AI companies focus their first efforts on long-haul trucking.

It is a great use case for AI technology.

A competitor, which currently is enjoying more success than Embark Trucks, is Plus AI (Plus).

Currently, they are developing (and delivering) a level 2 solution they call a Highly Automated Driving solution - HAD.

They are also working on level 3/level 4 solutions, a long-term problem, while the HAD is the “here and now” solution.

Without comparison, Plus AI's HAD system might be similar to “the feel” of Tesla’s autopilot.

In Norway, where I am originally from, “everyone” drives a Tesla. A friend of mine always talks about how he loves his autopilot. From what he has told me, the PlusDrive autonomous driving technology sounds similar.

(Sorry in advance to Plus AI if I am entirely off here).

Plus AI’s current (level 2) AI solution is called PlusDrive. It is a “perception system” with deep learning models to quickly, accurately, and safely perceive the vehicle’s surroundings, predict what is coming next, control the truck, and take the necessary actions, like stopping, slowing down, changing lanes, etc.

The trucks using Plus’ solution have the same technology stack as Embark Trucks without the intent of being fully autonomous: Radar, Lidars, GPS, cameras, processors, and an AI processing millions of data points in real-time.

Plus AI’s solution is also platform agnostic and can be used with most trucks.

The company is working on the level 3/level 4 autonomous system, which will be available once the systems and legislation are ready.

Until then, all self-driving solutions need rigorous testing and development.

Only when it has proven safe in all situations can it be released “into the wild.”

But there are more players in the market than Embark and Plus.

Several have gone bust, while others, like Ford- and Volkswagen-backed Argo AI ($4 billion), have shut down their operations (2022).

TuSimple was the first company to successfully demonstrate a driverless pilot in the USA, covering 80 miles from Tucson, Arizona, to Phenix.

Waymo is a Google company that has been testing its self-driving technology in California and Arizona for a while. In June 2022, they announced a partnership with Uber Freight to accelerate the future of logistics.

Daimler Trucks is the largest truck manufacturer in the world, and naturally, they are working on level 4 solutions for their range of trucks.

Here is a video from Daimler about their trucks and AI technology. Don’t mind the CEO sounding like a robot, put intended, and the highly scripted and “salesy” nature of the video:

And least not forget the joker, “the elephant in the room,” Tesla…

The company is scheduled to launch the Tesla Semi in 2023.

It will be a level 2 truck, most likely level 3 ready.

One thing is sure; the world will know about it when it launches, as the media will be in a frenzy.

Implications of Artificial Intelligence in trucking

So what will be the consequences and the implications of AI technology entering the trucking industry?

Is it all doom and gloom, or will it offer benefits and have a positive impact?

Will the $14 trillion industry be left in ruins by the introduction of robots and self-driving trucks?

I don’t think so!

Will it change the way it is today?

Absolutely!

At the moment, self-driving trucks are, as you can see from the Embark and Daimler videos, sold as solutions to two problems: 1) staff shortages and 2) road safety/accident-free driving.

Embark Trucks goes as far as to say that with the introduction of their solutions, the most straining and challenging part of trucking can be eliminated by automation.

At the same time, the resources (truckers) can be relocated to first- and last-mile driving.

They also claim that this increased productivity will create more jobs in other areas of the industry, typically the first/last mile driving.

Personally, I think this holds true in the shorter term, while I think long-term AI will eliminate jobs.

Why?

Because all car companies seem to want self-driving cars.

It’s the future!

Tesla is the leader in this space.

It is just a matter of time before level 4 autonomous cars become a thing.

Google, Apple, Amazon, Tesla, and many more are working on solutions.

Self-driving buses and taxis are the first to be disrupted, but why will not solutions for logistics be developed too?

Off course, it will!

In Norway, my home country, the company Aviant is building and deploying delivery drones for remote areas.

While in Tallinn, where Practical AI is based, they use delivery robots for food deliveries.

Deliveries are the “plumbing” of the economy, and it will be “attacked” by self-driving solutions.

I think AI companies can achieve economies of scale in the last-mile industry due to its sheer size.

Therefore, this part of the logistics industry will be disrupted eventually.

Regarding road safety and AI, Daimler Trucks says enhanced highway driving brings them closer to their vision of accident-free driving.

Humans are full of faults, and road accidents affect millions yearly.

We get distracted, are tired, stressed, emotional, hungry, and whatnot.

This affects our perception and focus and, thus, how we drive.

AI technology and self-driving trucks have 360 degrees of vision and can sense things around them, make advanced calculations in milliseconds, and act and react accordingly.

We only have two eyes and can observe in one direction at a time.

AI technology in trucks and cars clearly enhances transportation and, once tuned, will increase safety on the road considerably.

I genuinely believe so.

However, 100% accident-free roads are a goal we can strive towards but a reality we will never reach.

The complexities of roads in cities and urban areas, weather conditions, visibility, snow, ice, light, sun, darkness, tunnels, etc., all pose challenges humans and self-driving tech face that increase the risks of accidents.

The bigger problem self-driving trucks are trying to capitalize on - I mean the problem they try to solve - is the trucker shortages.

Trucker shortages are a real problem for the industry.

The constant need for deliveries (convenience), coupled with a decline in recruitment and increased average age of truckers/retirement of truckers, sets the stage for a perfect storm.

By 2030, America will be short of 160,000 truckers and more than double that in the EU, at around 400,000.

Artificial intelligence in self-driving trucks is one way to solve this problem.

But the industry risk becoming a victim of this success, as AI systems are easy to replicate.

The efficiency will be attractive to business owners. The self-driving truck does not need breaks, pee stops, food breaks, sleep breaks, etc.

An autonomous system can just keep on truckin'… pun intended.

While researching this newsletter, this video of an interview (with two controversial figures) appeared in my Youtube feed. Take politics out of it; does Tucker Carlson have a point?

Here is the link to the video. Somehow, I could not embed this video…

There is also the case of self-driving trucks being more sustainable and environmentally friendly.

The trucks can work when traffic is low, like in the evenings and nights, and thus drive more efficiently using less fuel and not adding to the congestion and pollution in urban areas as much.

Tesla Semi will even eliminate pollution altogether.

Long-haul trucking is inevitable, and from a pure efficiency perspective, I see this as a great use case for AI and self-driving trucks.

In other words, self-driving technology can utilize a truck 24/7, going from one shift per 24 hours to three.

If that is not productivity gains, I don’t know what is…

If it is the politically sensible thing to do, is a whole other question (ref the Tucker Carlson video above).

Will artificial intelligence replace truckers any time soon?

I don’t think so!

But changes are coming, and the logistics and trucking industry needs to be ready.

It will be here before they know it…

AI art of the week - the future of transportation

Image created using generative AI Lexica Art.

DuckDuckGo jumps on the generative AI craze 

Internet’s favorite privacy search engine DuckDuckGo has beta-launched DuckAssist, a generative AI assistant utilizing “natural language technology” from both OpenAI and Antrophic.

I find their approach to adopting AI into their search engine interesting.

DuckAssist utilizes one source to create its summaries: Wikipedia.

DuckDuckGo chose to work with Wikipedia because it is a public resource with a transparent editorial process that cites all the sources in an article.

Since Wikipedia is frequently updated, DuckAssist will deliver fairly up-to-date information.

The tool will occasionally pull data from Encyclopedia Britannica, but Wikipedia is the primary source of information.

ChatGPT, as you may know, pulls a lot of its information from Reddit, an open network of communities on various topics, and a wide range of other sources. However, Reddit is the biggest stand-alone source for ChatGPT.

Using one or two sources helps eliminate “hallucinations,” which generative AI and ChatGPT suffer from.

The tool is only available via their app for now. However, it is expected to roll out fully coming week(s).

I did have an opportunity to try it out.

It looks pretty cool.

You will get the DuckAssist prompt above the search results page when you search. Then you can press the “Ask” button and get a summary written by the AI with Wikipedia as the source.

The output looks like this:

If you want to try it out, you must download the app.

If you do not wish to download anything, you can go to their homepage and try it out from time to time, and once it is implemented, try it out that way.

Here is the DuckDuckGo search engine: https://duckduckgo.com/

AI experiment of the week - ChatGPT helps Data scientists

I came across a Youtube video earlier this week by Ken Jee called “How I use ChatGPT as a Data Scientist (4-5h saved per week).

That’s remarkable, I thought.

In short, Ken shows us how he saves a lot of time each week coding and helping him save (or gain) time in his work as a data scientist.

Have a look for yourself.

It may be a bit “nerdy” - it is programming, after all.

I find it fascinating how AI can assist us in so many aspects of work and life.

This is just one more real-life example!

Quick bites - interesting AI news

A CNN reporter plays around with voice AI and calls his parents to see if they would notice. It is about a 5-minute video worth a watch.

Bing, Microsoft's search engine with integrated AI Chat capabilities, crossed 100 million daily active users for the first time in history. Surely Microsoft has done a scoop with OpenAI against Google!

If you found my previous newsletter about AI in food production interesting, MarketWatch made a 3 min video about the same topic. It's pretty solid. I cannot but wonder, though, if any of their journalists are subscribed to this newsletter (let me know if you are 😉).

NASA uses AI to design hardware that is up to 3x better in performance/stronger than their human-developed counterparts.

And that is all for this edition.

If you have been subscribing since the beginning of the Practical AI newsletter, you may notice slight changes.

The newsletter is evolving, and I am starting to find a form I like.

There are many great AI newsletters out there; most, however, focus on curating the latest and greatest in AI.

Personally, I find the rapid development of AI overwhelming and impossible to keep up with.

The exciting part of AI, to me, is how it transforms the world around us.

New AI tools are cool and all!

But what I am trying to do here is to pull the curtains aside and explore what the AI revolution actually means for businesses, work, and our future.

In other words, I will continue doing one deep dive a week but with fewer sections about “everything else” going on in AI.

Other newsletters handle that part much better than I ever would.

I appreciate you being here, and I hope you find these write-ups helpful or interesting and that they give you food for thought.

If you have any questions about AI or any feedback, just hit reply or tweet me @thomassorheim 

AI’ll be back.

PS! 
Why did the data scientist refuse to go on a date with a statistician?

.

.

.

Because they couldn't find any correlation between them.

This is not the end. It is where the fun begins!