I think most people believe that the problem with robots is that we don't have the right software, and if we just knew how to program them then today's robots could be incredibly useful in everyday life. From that perspective, this move from OpenAI seems dumb.<p>That belief is wrong. Today's robots can't be made useful in everyday life no matter how advanced the software. The hardware is too inflexible, too unreliable, too fragile, too rigid, too heavy, too dangerous, too expensive, too slow.<p>In the past the software and hardware were equally bad, but today machine learning is advancing like crazy, while the hardware is improving at a snail's pace in comparison. Solving robotics is now a hardware problem, not a software problem. When the hardware is ready, the software will be comparatively easy to develop. Without the right hardware, you can't develop the appropriate software.<p>OpenAI is right to ignore robotics for now. It's a job for companies with a hardware focus, for at least the next decade.
Makes sense I guess, integrating robot hardware requires an entirely different set of skills to ML research and has a much slower dev cycle.<p>I think OpenAI has progressively narrowed down its core competency - for a company like 3M it would be something like "applying coatings to substrates", and for OpenAI it's more like "applying transformers to different domains".<p>It seems like most of their high-impact stuff is basically a big transformer: GPT-x, copilot, image gpt, DALL-E, CLIP, jukebox, musenet<p>their RL and gan/diffusion stuff bucks the trend, but I'm sure we'll see transformers show up in those domains as well.
Fascinating in the wake of Fei Fei Li's lab publishing significant work on embodied intelligence...<p><a href="https://arxiv.org/abs/2102.02202" rel="nofollow">https://arxiv.org/abs/2102.02202</a><p>Not to mention a bunch of relatively inexpensive reinforcement learning research relying on consumer knockoffs of Spot from Boston Dynamics...<p>Really does seem like they are following the money and while there's nothing wrong with that it's also nothing like their original mission.
Is the prevailing opinion that progress in reinforcement learning is dependent on algorithmic advances, as opposed to simply scaling existing algorithms? If that is the case, I could see this decision as an acknowledgement that they are not well positioned to push the frontier of reinforcement learning - at least not compared to any other academic or industry lab. Where they have seen success, and the direction it seems they are consolidating their focus, is in scaling up existing algorithms with larger networks and larger datasets. Generative modeling and self supervised learning seem more amenable to this engineering-first approach, so it seems prudent for them to concentrate their efforts in these areas.
This is lunacy. The first country/company to replace human labour with general bipedal robots, will reap wealth beyond imagination. The short sitedness is astonishing, if you ask me.<p>I genuinely believe how we as a society act once human labour is replaced is first aspect of the great filter.
Interesting contrast to another story today: <a href="https://ai.googleblog.com/2021/07/speeding-up-reinforcement-learning-with.html" rel="nofollow">https://ai.googleblog.com/2021/07/speeding-up-reinforcement-...</a>
I think the comments are confounding shutting down the robotics research team with eliminating all RL research. Most robotics teams don't use data-hungry RL algorithms because the cost of interacting with the environment is so expensive. And even if the team has a simulator that can approximate the real world to produce infinite data, there is still the issue of the "simulator-gap" with the real world.<p>I don't work for openAI but I would guess they are going to keep working on RL (e.g hide and seek, gym, DoTA style Research) to push the algorithmic SoTA. But translating that into a physical robot interacting with the physical world is extremely difficult and a ways away.
Curious idea:<p>With the mentioning that they can shift their focus to domains with extensive data that they can build models of action with etc... Why not try the following (If easily possible)<p>---<p>Take all the objects on the various 3D warehouses (thingiverse, and all the other 3d modeling repos out there) -- and have a system whereby an OpenAI 'Robotics' platform can virtually learn to manipulate and control a 3D model (solidworks/blender/whatever) and learn how to operate it.<p>It would be amazing to have an AI robotics platform where you feed it various 3D files of real/planned/designed machines, and have it understand the actual constituancy of the components involved, then learn its degrees of motion limits, or servo inputs etc... and then learn to drive the device.<p>Then, give it various other machines which share component types, built into any multitude of devices - and have it eval the model for familiar gears, worm-screws, servos, motors, etc... and have it figure out how to output the controller code to run an actual physically built out device.<p>Let it go through thousands of 3D models of things and build a library of common code that can be used to run those components when found in any design....<p>Then you couple that code with Copilot and allow for people to have a codebase for controlling such based on what OpenAI has already learned....<p>As Copilot is already built using a partnership with OpenAI...
I'm sure the overhead and upkeep of a robotics lab far outweighs that of a computer lab for software research.<p>Are there any Open* organizations for robotics that could perhaps fill the void here? I think robotics is really important and I think the software is a big deal also, but it's important that actual physical trials of these AIs are pursued. I would think that seeing something in real space like that offers an unparalleled insight for expert observers.<p>I remember the first time I ever orchestrated a DB failover routine, my boss took me into the server room when it was scheduled on the testing cluster. Hearing all the machines spin up and the hard drives start humming, that was a powerful and visceral moment for me and really crystallized what seemed like importance about my job.
Designing robots to pick fruit and make coffee / pizzas cannot have a positive ROI until labor laws make the bsuiness-case for them. Majority of use cases where we can use robots for activities humans cannot perform (fast spot welding on production line, moving nuclear fuel rods, etc) have been solved. It is smart to focus on language and information processing, given that we are producing so much more of it, everyday.