Johns Hopkins Beast | Hacker News Reader

Johns Hopkins Beast

en.wikipedia.org

198 pointsPetiver48 comments

Video: https://www.youtube.com/watch?v=zMpEeag7kkM ("APL Flashback: 1965 - The "Hopkins Beast" Mobile Automaton ", taken from the 2nd external link in the Wikipedia article, of the Mod II)

also https://www.youtube.com/watch?v=szGFt28BiBY ("APL Flashback: 1964 - "Ferdinand" the Mobile Automaton", a similar device)

Wow, sorry for diverting the topic a tiny bit but there's a strong star wars IV reminiscence in these vids. I wonder how common this robots were in the minds of people of the 60s/70s.

Amazing how my generic Roomba does all those same behaviors (navigate a room with obstacles, find its own charger) shown in that first video, but it took another 60 years for such a thing to be a cheap consumer product.

Wonder how long it will take to find a Roomba that can be scheduled to run more than once every 24 hours? /J

It is much cheaper to make a prototype than making something at scale (although publishing the circuit diagrams and so forth might have been interesting. The biggest problem in this would be having a battery that would power a Roomba to do vacuuming without looking for an outlet every five minutes. Roombas also have edge detection so it doesn't fall down the stairs. But, all of the mother of all demo's technology took forty years to implement also.

> Roombas also have edge detection so it doesn't fall down the stairs.

From article, “It could also ultrasonically recognize the stairway and doorways to take appropriate action.”

Hi all, I work at JHU apl in the r&d department and wanted to say that we're hiring right now!

Pay isn't FAANG level but you get to do some really interesting work and you have a good bit of freedom in choosing what projects you get to work on. If you're a curious person that likes to constantly keep learning to tackle new challenges, it's a great place to work.

That being said, our hiring process is ancient and takes a good bit of time to go through and you'll probably need to be able to get a clearance of some sort.

Probably implied, but I’m assuming this requires relocating to Baltimore?

It's located between Baltimore and DC in Laurel Maryland. I commute from central baltimore and the drive is around 40 minutes.

It's looking like some amount of in office time will be required (2-3 days a week) although some people have been working remotely since before the pandemic. If you're working on classified stuff you'll also need to be on campus.

> probably need to be able to get a clearance of some sort.

Why is this the case? Is everything being worked on government funded? I would probably apply otherwise.

Yes; APL is a University Affiliated Research Center (UARC) and while we sometimes do work with industry, our whole purpose is to support the US Government. That doesn't necessarily mean you need a clearance, but much of our work is national security related.

https://en.wikipedia.org/wiki/University_Affiliated_Research...

Yeah, I had no idea that was the case. No interest in building missiles or weapons anyway.

They make bombs, crazy spy stuff, crack crypto. Def needs clearance.

Any advice on getting through HR?

I had a great time visiting APL in 2016 or 2017, when I had nothing but a preprint and a prayer. Funding fell through for that job, but I've applied to several others. I feel like I'm a much stronger candidate now, but subsequent applications have just vanished into the ether without even a phone screen.

It really depends on what job you’re going for. There are seven big sectors with tons of variation in culture and processes between them. I’m actually trying to find experience full stack devs right now, so if that’s up your alley we can definitely talk. If you’re looking for something else, I can for sure point your resume toward the right department.

Thanks! I was looking at REDD--my background is machine learning + computational/experimental neuroscience.

My email is in my profile, feel free to shoot me a resume and I'll pass it along!

Definitely reach out — neuro-AI group is actively hiring :)

Thanks! I'll shoot you and NavBaker an email this afternoon.

Nice to see fellow APLers here! I have been at APL 5 years and like it a lot.

> Johns Hopkins University Applied Physics Laboratory

I thought it was APL the language :)

I worked at APL for 3 years (2015 - 2018) and loved my time there. Although, the work got repetitive over time and was no longer challenging, the thing I loved the most was the social environment. Theres a lot to be said for having a good team/group of people around you. I still keep in touch with some of them.

A curious insight into the visions of future at the time. The robot [0] has curious resemblance to Daleks [1] of Doctor Who fame, with Daleks making their first on-screen appearance only a year or two after the initial release of this robot. Life imitates art imitates life.

[0] http://cyberneticzoo.com/cyberneticanimals/1962-5-hopkins-be...

[1] https://en.wikipedia.org/wiki/Dalek

Daleks, canonically, have a tentacled bio-monstrosity riding around inside.

Yeah just like any other university project.

My eyes popped when I learned it functioned with just some basic transistors and could find outlets and recharge itself. Amazing!

Is there a surviving photograph?

A sibling comment uncovered not just a photo but a whole video demo: https://news.ycombinator.com/item?id=30861972

this thing reminds me of the [something] paradox. humans maximize what they can use, not what they're able to do with it, unless forced to. modern day roombas have much more computing power than most (any?) computers did at the the time the Beast was created, but it performs essentially the same function. you could make the argument that the humans making the roomba spent less time/energy/money developing it with modern computing tech and high level languages than researchers did developing the Beast, but the apparent requirement of an increase of computing power by many orders of magnitude for a reduction of effort by one or two orders of magnitude is completely insane.

Yes, see: software bloat.

Hardware constraints force good engineering. The fact is that ever growing hardware capabilities has enabled poor engineering practices to succeed. The technical debt is already being exposed in infosec / related supply chain issues.

During the 1940's: https://en.wikipedia.org/wiki/William_Grey_Walter#Robots

I see this as an early example of BEAM robotics[0].

[0]: https://en.wikipedia.org/wiki/BEAM_robotics

I wonder what people who built this with dozens of transistors would make of the fact that we can't really do a lot better with billions of them, a radio link to trillions more and 60 years of work.

Maybe Boston Dynamics has indeed done a lot better than The John Hopkins Beast

I'd say the Internet is a good example of a cybernetic model that is a step up in scale from this. That's pretty impressive!

We actually do a lot of things much better than this lil guy, that just don't take the form of robots the way the futurists and scifi authors of the 60s predicted. I can order food online to be delivered to me on a device the size of my hand, that also connects me to most of the earth's population, and gives me a direct line to all of humanity's knowledge, art, and perspectives. We really do live in the future, it just doesn't look like the Jetsons!

Well, most of those billions of transistors are memory rather than logic circuits. (And no, 640K is no longer enough for everybody, unfortunately.)

Are you sure? My roomba can self navigate the house, and go back to a charging station as needed - including if the room geometry changes around it or if it's dropped in an unknown area.

They would chuckle when they realized how instead of seeking to replace $8/hr labor to do simple tasks with fancy robots, we instead decided to focus on telecommunications and mobile computing, and now anyone can buy a smartphone for $800 that has a 120 fps HD+ screen (inconceivable to someone in the 60's) that can download 100mbps (inconceivable) almost anywhere in the world (inconceivable), record 4k/60FPS HDR video(inconceivable), store 1TB data (inconceivable) buy anything in the world and have it delivered tomorrow (inconceivable). We've built far more interesting and useful things.

Furthermore, It's not that we can't, we just don't want to. We like interacting with the world around us. Most people enjoy cooking, cleaning our living spaces, etc. As it ties us to the world. If we outsourced this to robots, we would maybe save some time, but then have to deal with robot repairs and troubleshooting.

Which would you rather do: Make dinner on a lovely Le Creuset pan or troubleshoot your Super Cook-a-tron 3000 because it broke down again?

> Furthermore, It's not that we can't, we just don't want to. We like interacting with the world around us. Most people enjoy cooking, cleaning our living spaces, etc.

I'm not sure. From what I've seen, cooking and cleaning are usually considered to be inconvenient. I'd wager that most people do these things themselves because the alternatives are expensive in the long run, in more than one way.

Seems like you’re discounting the reliability of a hypothetical Super Cook-a-tron 3000 in furtherance of a particular native.

Professional chefs and housekeepers exist.

We have amazing industrial robotics now. We also have bots that can navigate real world organic spaces, they're just expensive and nobody knows what to do with them in any consumer appplicationn.

I think they would marvel at the scale of industrial automation, assembly lines in factories, robots packing and moving boxes in Amazon warehouses, Roombas in people's homes, autonomous cars on the roads and drones in the sky, all the stuff Boston Dynamics comes up with.

this thing sounds smart as, if not smarter than, a roomba.

This thing probably cost 1000x a roomba to build. Current supply chain issues aside this ‘future’ tech is now accessible to most of the population now.

A roomba builds a map of its space and wont go over cliffs.

We're doing OK though: https://www.youtube.com/watch?v=fn3KWM1kuAw

Dunno - did you see that Waymo is operating autonomously in SF now? I think that the revelation of the complexity of the real world is the most interesting part of the story.

> 40 kilocyles

First time I hear that term. It hurts to hear it...

The Hopkins Beast was so far ahead of its time.

early roomba

Reply on news.ycombinator.com