Actually one more thing on the gym part. Most have large mirrors that go all the way down. These utterly confuse lidar guidance. The easy solution is to cover the the bottom part of the mirrors, so that lidar sees it as an obstacle, rather than thinking that there's free space there going all the way to where it reflects the beam.

Firmware updates after "basic functionality is robust" are nonexistent in my experience. For everything but maybe the most expensive modules.

Notably it's commonly recommended that you use double sided tape for small rugs that get moved around by the vacuum. They cost pennies at pretty much any website selling accessories for robotic vacuums.

No idea what's up with your gym equipment, but general recommendation is that you put as much of the stuff strewn around the floor on mounts, chairs, tables etc. Floor should be reasonably obstacle free. You'll find this information in the manual.

If instead of "The internet", whatever that means, you go to pretty much any store that sells large amounts of robotic vacuum cleaners, you'll quickly notice that all of the most popular models sold are lidar models. Because of just how fundamentally awful the early non-lidar models are in navigation.

It's easy to tell because lidar versions have the lidar tower on top.

If I remember correctly, consumer models tend to have very limited range of LIDAR in the tower on purpose for large manufacturers that have separate consumer and industrial models. It prevents consumer versions from being used in large commercial spaces instead of much more costly industrial versions.

It's a product segmentation issue. You can generally get around it by putting some object in very wide open spaces that lidar can track within its limited range.

>doesn't make mistakes that a child could avoid

If you look at my post history on this subject, you'll find that not only am I aware of this issue, but I even explained in great detail the theory behind why it makes those mistakes, and what it will take to eliminate them.

Specifically, the problem is embodiment issue. Unlike luxury belief of very rich and decadent Westerners dictate, we are not a "ghost in the shell", the consciousness in a machine. We are an integrated whole, with consciousness being a sum of all bodily inputs. It's why Ozempic works to prevent obesity. It generates a behavioral change by affecting specific transmitters going from gut microbiome to brain, without ever touching the brain, the supposed location of "the ghost".

AI is in fact a ghost in the shell at this point. It has no contact with reality other than what we artificially feed it. As a result, it has no context of what it does. To grant it ability to grasp context, we need to embody it and allow it to learn like a child would.

Modern ones haven't built a map visually in a long time (if ever). They build it from sensor fusion of lidar (mounted in lidar tower on top of the unit), IR mapping using grid projector and camera mounted in front and on the right side, and finally kinetic bump sensor in front and on the sides.

Maybe some original iRobot built on basis of visual mapping.

It's probably not known yet, and will be figured out during the bankruptcy proceedings.

Mine automatically goes to higher suction levels on carpets (it recognizes them automatically, and I can set it manually in the map too if I want), and can cross pretty much all thresholds.

There are actual ramps designed for robotic vacuums if you have ramps or shallow steps so it can cross them. They're really cheap too.

Higher end units have rotating mops that can deal with mud, and they also now have forward facing cameras with object recognition that works quite well on things like animal excrement.

No idea how chairs would even be an obstacle, as long as the bottom is high enough off the floor that robovacuum's lidar tower fits under it.

I wonder if I'm much cleaner than I thought I am, because I'm yet to encouter problems with my low end cleaner. It sucks up everything, it has pretty much the lowest end mop variant that is good enough due to the way vacuum's routing will drag it over every spot at least twice.

Pretty much the only thing that needs attention after it's done is corners and a few places with plants and wiring on the floor that are set as no-go zones.

There are several PRC companies that have overtaken iRobot in this field. This is in fact one of the main reasons for its decline. You get a lesser product for more money because of the brand, and eventually customers left for better and cheaper options.

This field is actually a really good example of the "US invents, China iterates and EU regulates". US invented the robotic vacuum, but PRC iterated on it and made it much better. Hence iRobot's primary supplier being from Shenzhen.

This doesn't make geothermal universally available or anything of the sort.

It just finds a handful of locations we missed. It's a specialist application. There are similar applications across countless fields. To make a connection to reality, it's why nvidia rtx 5090 are selling for way more than MSRP. They have enough VRAM to fit narrow models for training purposes. So every research institution is vacuuming them to run such models to do inference on their data searching for things we missed.

Large scale LLM models on the other hand are universally applicable, which is why they are the buzz. It's what people actually can use in everyday life without being a narrow specialist.

You remind me of the redditors who crusade against AI art, who have been recently posting about how they're winning, because there's not as much AI art any more.

The level of cognitive blindness there is funny, just as it is here.

These uses have been massive for a while, as foundational principles are exactly the same as ones in LLM. It's just that where LLM addresses relational likelihood of letters, words, sentences and paragraphs, these AI's address relational likelihood of various geological observations.

Other useful applications you don't hear about if you don't follow it include everything from medical imaging and diagnosis to searching for new sources of oil, to discovering new antibiotics.

It all works on the same principle. The difference is specifics of model and training materials.

To be fair, Opera the company behind the browser changed hands several times since the original Opera. At which point it became a barely changed Chromium skin.

If you want a browser built by team formed from people that made original Opera with the kind of UI design philosophy that made original Opera's fame, you don't use modern Opera. You use Vivaldi. Modern Opera doesn't really have any meaningful commonalities with original Opera. Engine is different, team is different, even design is different.

Vivaldi at least maintained focus on specific UI things like panels that Opera was known for. It's in fact one part of the browser that isn't open source. They have proprietary UI that sits on top of Chromium.

Has it ever occurred to you that comparing any modern Western leader to North Korean leader as similar is so obviously wrong that you should stop and reconsider everything you believe about real world?