Science in Progress: The Self-Driving DeLorean

Welcome to the debut episode of Science in Progress! Tested’s Kishore Hari and Indre Viskontas explore the labs where scientists invent custom tools and technologies to further their research. Their first visit takes them to a workshop where Stanford engineers study autonomous driving by turning a DeLorean into a self-drifting car!

Comments (30)

30 thoughts on “Science in Progress: The Self-Driving DeLorean

  1. This is awesome, I can’t wait to see more of this series.

    By the way, the name tag that pops up at 9:11 says “Jonathan Goh: Price and Description Here”

  2. Awesome! Working in science myself I was hoping for such a thing. 🙂

    And ps; I’m so jealous of those LHC pictures at CERN! I was there just when it was shutdown, so I could’ve visited it down below, but a very annoying scientific group decided to do a last run. grrrr 🙂 There are so many groups there that it’s almost a battle of who gets time on it.

  3. First off, a big warm Tested community welcome to Dr. Indre Viskontas! I absolutely love how so many of my favorite makers, podcasters & YouTubers are converging. Yesterday it was Laura Kampf in Adam’s cave and now Indre Viskontas as a part of the Tested team. If you can get Jimmy DiResta to do a 1 Day Build throw-down again Adam, my mind will explode! 🙂

    I was first introduced to Dr. Viskontas thru her lecture series from the Teaching Company’s Great Courses (12 Essential Scientific Concepts). Then a couple years later, I saw her interview Adam for Inquiring Minds (which is where I discovered the equally amazing Kishore Hari). Thanks again Tested for continuing to expand your boundaries. You guys are always my first stop on the Internet when I fire up my browser.

    A question for Adam – at the 3:20 mark, you refer to getting in trouble for drifting. Is this the incident that was alluded to in (and excised from) the Still Untitled episode (1/27/2015 The Austin Trip) where you talked about Tom Sach’s Chevy Caprice?

    Definitely keep it coming with this series – I’m eagerly anticipating the next one!

  4. this looks great! i’m super stoked for more episodes, and hello dr. indre!

    re the morality question: the delorean is super cool, but pretty much a one-trick pony, right? so if all it’s going to do is master one task, like drifting, that’s codeable in a more or less straightforward manner. as would be a car that can do a limited number of tricks. but to tackle the endless variety of emerging situations in real world traffic, i’m guessing that it isn’t a possibility anymore to code a routine for tackling each type of situation. that’s just too many situations and variants and subvariants, even if each routine has a certain span for corrections.

    so if one were then to decide to tackle this with machine learning, the problem becomes even more tangled. if someone wrote a routine that prescribes action X for input Y, and input Y could be faultily triggered at the wrong time, programmer error still is a kind of analogue situation to driver error.

    but when you employ machine learning, you have much less direct input on what signals get processed and lead to certain outcomes. so who caused the driver error, then? even when those models get subjected to a comprehensive test suite and pass it with a certain rate of correct decisions, we lack a responsible entity that caused a faulty decision.

    as impressive as machine learning results can be – at uni, a fellow research group is OCRing medieval manuscripts with a machine-learned model trained with just 2–3 pages of transcribed manuscript, at correctness rates over 98%, and we’re about to try training it to OCR arabic – it’s both the tool of choice for approximating correct interpretations based on highly variable input data, and also by design somewhat opaque and not man-made.

  5. Thanks Eric! I’m super excited about this show and feel a bit like a kid in a candy shop deciding which treat to try next…

  6. Kishore and Indre made me sign up for premium. After 17min I actually thought “Awe… it’s over?” Can’t wait for more episodes!

  7. I am very much looking forward to the coming episodes of this show. I really like the concept of exploring the science and technology that is still under development and I think that the show is of to a good start with this episode.

    I do feel that the dialog between Indre and Kishore seemed scripted or forced, given the casual setting.

  8. This was a great first episode, I hope you get to join the action next time, though that beer did look delicious 😀

  9. This is what makes me nervous, precisely! Some unforeseen circumstances that lead the machine to make decisions that violate the do no harm clause. Then again, it is an interesting question whether that would be such a small proportion of the time, compared with the number of motor vehicle accidents that now result in fatalities. If we reduce the base rate of accidents to the point where they are truly rare, then once in a while, we have a problem, maybe humanity is better off on the whole.

    But I still don’t see that autonomous cars are a better solution than mass transport. Or teleportation… are we falling into the silicon valley over-engineering problem rather than finding radical solutions that fix the problem?

  10. count me in on team teleportation! self-driving cars optimise (citation needed) a detail in a set of options we already have. teleportation creates new ones. (it’s about time the real world caught up with the internet’s redefining of distance anyway.)

  11. Really happy to see this. Tested- any chance your premium video player will ever support Chromecast? I’d love to actually watch this on my television.

Leave a Reply

Your email address will not be published.

Show And Tell

Tested in 2020: Ariel’s Favorite Things!

Ariel brings in a recommendation for an accessible and power…

Podcast - Adam Savage Project

Greg Munson – The Adam Savage Project – 11/24/20

BattleBots is back this week! We're joined by the show's co-creator, executive producer, and robot combat veteren Greg Munson to talk about the process of producing a season of BattleBots during lockdown, the roster of competitors and robots appearing this year, and what how he and the BattleBots team are thinking about ways to evolve and…

Making

Adam Savage’s Favorite Tools: All About Awls!

In showing off the usefulness of a small awl for marking you…

Show And Tell

Tested in 2020: Sean’s Favorite Things!

Our resident 3D printing expert and modelmaker Sean checks i…

Show And Tell

2001: A Space Odyssey Moon-Watcher and Bone Replic…

Adam Savage shares two pieces of his collection from the fil…

Show And Tell

Tested in 2020: Bill’s Favorite Things!

Tested contributor Bill Doran checks in from his workshop to…

Culture

Filming Model Miniatures from Moon (2009)

Duncan Jones' Moon is one of our favorite science fiction fi…

Show And Tell

Tested in 2020: Mel’s Favorite Things!

Fabricator, engineer, and Adam's former shop assistant Mel s…

One Day Builds

Adam Savage’s One Day Builds: Luke’s Lightsaber!

Time to make a laser sword hilt! Adam takes a cylinder of al…

Show And Tell

Tested in 2020: Jen’s Favorite Things!

Jen's favorite things from this year include new fabrication…