The CIA’s Amazing Bots

When you have a virtually unlimited budget, you can pull off some amazing things. This has become most evident recently as the CIA has been showing off some of its old tech. That dragonfly you see above is near life-size and actually flies. They hired a watch maker to build a tiny internal combustion engine to run it. That alone is pretty amazing, but this thing was actually flying in the 70’s. Upon further inspection of the wings, we actually have no idea how this sucker is supposed to fly. Despite our skeptical viewpoint, you can see a tiny clip of it flying after the break.  You can also catch a video of “charlie” the robot catfish.

[via Botjunkie]



56 thoughts on “The CIA’s Amazing Bots

  1. A combustion engine in a dragonfly in the 70’s?! No way man… The smallest I know of is a 2-stroke .25cc and that is about 3x1x2cm in size alone. it needs glow-plugs, a starter-system (Just the flick of a finger against one of its wings?! come on…)
    Don’t tell me we’re all going to watch 3D tv’s just by rapidly blinking.

  2. Seismic Intruder Detection Device at

    I bought three of these (or something exactly like it) from a surplus place in the mid-70s. The thing poking out of the top left is the power switch, you break it off, pull out a little plastic nub and tear that off, then re-insert the part. It’s powered by a stack four mercury cells and has a seismic detector that looks like a BB on the end of a small spring, sitting inside something the size of a .22 shell. Every time it gets bumped hard enough, the BB contacts the metal cylinder and transmits (not a coded message) a burst of RF at around 50 MHz. You can break off another nub and insert a tuning tool to change its transmit frequency.

  3. That’s impressive although, at the time, I’m not sure why somebody would have selected that over just a really good long-distance microphone. The thing is so difficult to see, I can’t imagine flying it more than 10-20 yards away, much less around the corner.

  4. I guess it works the same way as the backup diesel generators in nuclear subs where there’s no crankshaft but the pistons are opposed and bounce off a cushion of air trapped between them. Some of the combustion pressure is bled off to drive turbines in those things, but with this I guess they would have linked the wings direct to the pistons and once it hits resonance it just keeps going. To get it started you would waggle the wings up and down with a separate starter.

  5. if the engine was a 2-stroke diesel, it could indeed by started by cycling it until the compression was high enough to cause self-ignition.

    some people are very skeptical of this, that it’s just propaganda (hey guys, look what we built 40 years ago!)

  6. I’m voting propaganda too

    First, here’s the original source for the quoted text in the linked article (it’s a reuters article from 2003…don’t know why it was just rediscovered yesterday):

    I have several issues with this, and I’m sure there’s plenty more to say about the “oscillating engine” and “fuel bladder”, but one that I would really like to prod first is the “steering mechanism”, a laser.
    1) Diode lasers weren’t even really developped until the mid 60’s. Diode lasers in the late 1960’s were only achieving continuous lasing in labs at ~80K. Once the ambient temperature went up to room temperature, the current density shot through the roof. This problem wasn’t solved until the early 70’s.
    2) If the CIA was going to develop portable, room temperature laser systems wouldn’t THAT be the amazing development? Who would care about a dumb robot dragonfly?
    3) Even assuming the lasers were developed and produced, how the hell would you steer equipment with them?!?! Have the bug optically track the dot or something? If so, you would need tracking circuits, advanced detectors on the bug, some mini servo motor to control a rudder(? don’t know what else to call it), all technology that I’m pretty sure wasn’t being miniaturized to this extent in 1970.
    4) Even if you say the laser was used to relay instructions rather than project a dot (so as not to produce detectable communications), that still doesn’t eliminate the need for any control circuitry or receiving circuitry.
    5) Electric power. No matter how you dice it, the bug would need some kind of electric circuitry on it. How are you going to power it? Don’t say “battery”, because if that were the case, WHY BUILD A MINI COMBUSTION ENGINE IN THE FIRST PLACE?!

  7. “Upon further inspection of the wings, we actually have no idea how this sucker is supposed to fly.”
    What’s unclear about it? It works the same way the FlyTech Dragonfly works, right? The front of the wings move up-and-down, and the rest follows, pushing air backwards.

  8. I also think it’s pretty funny that they specifically say a watchmaker built it. I assume this qualifies you to build mini engines because you….work with tiny stuff a lot? Or something? Dragonfly, engine, CIA, watchmaker, laser….this whole thing sounds like some kind of spy madlib.

  9. It could be like 30x actual size – though I that would be retarded and wouldn’t blend in well.

    If it is real size – I am extremely skeptical – flying insects have incredible power to weight ratios that I was under the impression humans had yet to attain.

  10. the beginning of the movie was animation, why couldn’t the video clip part be that too? and was it just me, or was it impossible to see what the flying object was, besides a dark spot on the screen? I’m not saying its impossible, but it looks a lot like Hollywood animations.

  11. The notion that it has a itty bitty internal combustion engine in it is a little silly, but it is reasonable that it’s a windup clockwork device. I would guess it was uncontrolled, had a range of 20-30 yards, and a very small payload. This would make it slightly useful. On the other hand, what do I know, I’m just some guy on the internet.

  12. @ AG12

    how do we know? the government could have nanobots distributed to every household in the USA. they could have brainwash stations all over the country, and they cold be monitoring the internet for posts like this.

    *robot grabs from behind*

  13. @alan dulles

    I don’t doubt that someone back then had the hair-brained scheme to build a miniature flying dragonfly spy robot (these are the same guys who thought pumping people full of LSD might lead to a reliable way to extract information). I’m just saying that there’s no way it worked. At the very least, the methods described in the linked were at best, grossly exaggerated. The CIA’s whole job is to pass around information (both real and fake). I imagine at the time, nothing would scare a commie more than having to worry every time a bug flew by about being…bugged.

    1. They stated it was a complete failure in the control department. I don’t doubt a basic ornithopter that size can be built, but the picture of the wings shows no mechanical means of flapping. It all looks like one single piece. There is a high probability that I’m just missing something.

  14. Exactly, anyone could be anywhere in technological evolution. Ever inhaled a bug? It was full of nanobots.

    They don’t even have to build anything. They just have to engineer bacteria to perform tasks or transmit information with little protein or mineral antennas they would bio-accumulate once programmed to.

  15. I think it flew, but only a few feet and could not be directed or be used to spy at that point, it’s one of those concept with 30 components and they started on 1 of those components and that wasn’t even 15% done.

  16. There is a nice paper in the current Chemical Engineering Science Journal ( laying out why it is unfeasible to build an ICE smaller than 0.4cc currently. Apart from that, why should anyone in the 70s have built something like that, other than as a toy? RC was not good enough at that scale to do anything useful, so the bug would have been confined to preconfigured, line-of-sight trajectories. And know what other CIA tech can achieve that? A throw by hand.

  17. Engine probably the least problem needed to be solve here, to fly even just to freeze in mid air it need controls read more electric engines and power source. Aerodynamically it hard to control bugs they are not gliders and use turbulent propulsion which is chaotic in movement and need constant correction read uC

  18. I thoroughly recommend taking a look at and getting a copy of the book. I’m only a third of the way through and it seems the CIA had digital cameras back in the late 80s!

    Fascinating stuff. As for the Dragonfly the size of it and the noise the wings would make in flight negate any possible advantages you would gain from it. Something better would be a centipede or snake type design.

  19. Why not put in the “bug” inside of an actual dead bug, like a roach or beetle and throw it where you want? This fake machine they claim to have used doesn’t have any control mechanism. But more importantly, it has no stability mechanisms. Most flying devices rely on natural stability provided by properly designed wings and a center of gravity forward of the center of lift to fly without becoming unstable and falling out of the sky. This “device”, without active stabilization and without an understanding of the mechanics of vortex shedding, cannot possibly fly.

  20. The point is, as @Whatnot says, that this is merely one piece of a puzzle, a Proof of Concept if you will.
    Once you have the delivery mechanism, then you can work on the payload. Maybe we haven’t seen the payload yet, or it was never built as the delivery mechanism wasn’t viable.
    Either way, someone built this thing and it’s a great piece of tech from 30+ years ago, and shows that some people are way ahead of their time!

  21. If you look at the original CNN story carefully, you’ll notice it never says COMBUSTION engine, it says OSCILLATING engine. It could be a wobbler-type air engine, working off compressed air or a low boiling point liquid (butane? liquid nitrogen?) in the “fuel bladder”.

    Don’t get me wrong, I am almost entirely convinced the dragonfly thing isn’t true, but it’s not quite as far out as the story summary implies.

  22. Yeah. Clearly the insectothopter thing is not only old news (I seem to remember seeing coverage of it back as far as 2007 or so) but also blatantly false. It’s obvious to anyone with half a brain that this sort of crude technology demo couldn’t possibly have been developed in the 70s. In exactly the same way that it’s obvious to anyone with half a brain that sending people into space and eventually to the moon couldn’t possibly have happened in the 60s.

    To those with a whole cerebral orb, however, it may well look entirely do-able.

  23. No bigger than 0.3cc? Carp. The author of that paper has presumably not seen the *commercially available*, *mass produced*, 0.16cc Cox TeeDee engine. That’s a full-scale reciprocating 2 stroke gas engine rather than a simplified micro/nano power source, which could easily be made much smaller (for varying values of “easy”).

    Video of the TeeDee here :

  24. I agree the story is extremely suspicious, but I don’t know why so many people are assuming this is specifically propaganda. Who would this impress that wasn’t already impressed by men walking on the goddamn moon?

    Now on the other hand, the CIA put tons of resources into total red herrings intended to trick the soviets into wasting resources on wild goose chases, often involving insane multi-pronged deception. For example, there are known cases of plans for non-existent weapon systems being leaked to russian spies, while at the same time fake prototypes are “accidentally” left partially exposed to spy satellites. Every man-hour spent sifting the possibly true from the pile of obviously false was an hour not spent looking into the possibly true in more detail.

    Which seems more plausible: that the CIA thought this would be good propaganda, or that it was typical counter intel misdirection?

  25. actually this is feasible even with 1970’s tech.
    they developed basic control systems which you would think need an expensive (modern) fast micro but old fashioned analogue circuitry will work if you are very skilled in design.
    for example, a “chaotic” oscillator can be built with as few as two transistors and less than ten other components which can take information from a tilt sensor made from a vibrating mass bar and magnetic sensors and feed back into the system.
    Straight and level flight is simple, left/right and altitude feedback is also doable by altering the directional bias to the central pivot on the small resonating wing pair.

    As for steering during flight, any number of compact tunnel diode based RX circuits will work here and only need a handful of discretes.

    For once, 1970’s era parts are a boon for “black ops” because they are incredibly low power and can run from under a volt which is doable using simple thermoelectrics driven by the waste heat from the combustion chamber.


Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.