What's the best way/resource to get an honest/pragmatic view of where things stand with the "robots market" in general and how much and fast things are really progressing?
I remember seeing prototypes from Toshiba when I was 10 (20 years ago), and every few months, there is a company releasing an "amazing video." its mother company then spins it off like there's no adequate progress, and so on.
In it she covers the latest and greatest robot news, with occasional commentary/perspectives.
However to more directly answer your question, you need to know/talk to someone in the industry at the moment. I am not aware of a single “spot” that gives an honest in depth appraisal of where we are.
From my experience there is a ton of new “hardware” coming out, not just in the humanoid space (Agility Robotics being imho the most “real”), but also in lower cost robot arms, end effectors, sensors, and compute.
Where things are harder to track is where we really are in the software realm. If you look at software driving this hardware, most of it is early stages. Perhaps TRL level 3 to 5 at best. The higher TRL is non-intelligent control software (that is based on decades of work). The newer, AI/Machine Learning/“Smart” software tends to only have limited roll out. At best it will be a startup at the relatively early stages, but more often then not it is still a researcher sitting at a University or a large corporations research lab. In either of those cases, you will see single to at most double digit examples of those systems actually doing work.
However, to your point, it is super easy to create a single (or even a series) of cool videos… it just takes one success in 100s of takes. It is harder to make something that will perform day in and day out and really change the industry/world.
Lex Fridman has a long interview [1] with Marc Raibert, CEO of Boston Dynamics, which is really excellent. It might partially or wholly answer your question.
Talk to people in the area, I guess we do miss honest and straight forward source of info for the general public.
In general robotics flies under the radar because it's rare to see a unicorn or anything really flashy and there is a big gap between big aspirations and fake demos and real world applications with polished use cases and diligent design, processes, etc.
source: I'm a skeptic roboticist working in the industry.
I have zero ties to the industry.
Am I right to assume there's a lot of DoD-driven echo chamber? Material being produced for the big clients and contracts ?
I'm not based in the US to give you an accurate picture on this scene, most of it happens behind the curtains.
What I can say it's there has been always a movement to weaponize robotics in some way and this has gained interest from the market in the past few years specially with the Ukrainian and Palestinian wars. It takes time and a lot of money to polish an application like this, if there isn't a behemoth funding research and PD on this it will take a long time before it takes off, and I hope it never does.
> What's the best way/resource to get an honest/pragmatic view of where things stand with the "robots market" in general and how much and fast things are really progressing?
Like with every other market check if the product is available for sale and at what price point. And then look up what failure points people actually using the system are complaining about. (Because every system has problems and weaknesses. If you don't see reports about any then the system hasn't left the lab where the PR of it is controlled.)
worked examples:
washing machine (that's a robot alright, has a computer, actuators, sensors). Readily available commercially for 200-500 GBP. Usually works reliably, occasional reports of flooding the room.
robotic vacuum: Readily available commercially for 300-1k GBP. Works okay, reports about it spreading pet's poop around rooms.
spot from Boston Dynamics. Not as readily available as the above, but can be purchased. Reported price 74,500 USD[1] Seems to trip over its own legs sometimes in a hard to explain way: [2][3] (not to count as a dig against spot, seeing these issues is actually a great thing. It means third party people in the real world use it.)
atlas from Boston Dynamics. You can't buy it. No price advertised. You can't see third party reports of it malfunctioning. Not because it is perfect, but because nobody has access to it.
I like the IEEE Spectrum Video Friday blog posts. A regular fresh list of demos is not an overview, but it is an easy way to catch a glimpse of the long tail.
I'd say Tesla is the leader or could quickly become the leader given their intense investment in FSD. If a car software can "understand the physical world" using vision Ai / neural nets, it shouldn't be out of the question to reoptimize that software for the rest of the "physical world". Especially when you need a whole lot less safety standards compared to a 3,000lb 70MPH vehicle. Hell, the Optimus engineers said they were considering doing the first demo on a road since the software was so similar lol.
With FSD 12.3.3 released, it's clear FSD is getting smarter and smarter. How many of those releases left until people trust Optimus to fold their laundry? 1.0 Optimius will still be pretty dumb, but could still be worth the price (especially with continuous software upgrades!)
A road (most) has marked lanes and signage to provide a huge amount of contextual information. The world (and human interaction) is highly ambiguous and dynamic. Tesla is optimizing for the road.
I remember seeing prototypes from Toshiba when I was 10 (20 years ago), and every few months, there is a company releasing an "amazing video." its mother company then spins it off like there's no adequate progress, and so on.