Cooking with glasses(macwright.com)
macwright.com
Cooking with glasses
https://macwright.com/2025/09/21/cooking-with-glasses
14 comments
I agree with the general vibe of your post but let me make a tangeant:
> I always think of my Roomba in this kind of situation, namely when one tries to imagine a future where tech works outside of the computer. Basically my little vacuum robot mostly sits idling in the corner because... a basic broom or a normal vacuum cleaner is radically more efficient, even if it means I have to do so myself.
For me the whole value of a robot vacuum is that... I don't vacuum, the robot does, and when it does so it when I'm not at home. Reduce the dust periodically by 75% so that it's neater; I'll do a deep clean every few month or so.
(maybe what you're saying is that the cost of vacuuming for you is low enough that it's similar to launching the robot vacuum, in that case I understand you !)
> I always think of my Roomba in this kind of situation, namely when one tries to imagine a future where tech works outside of the computer. Basically my little vacuum robot mostly sits idling in the corner because... a basic broom or a normal vacuum cleaner is radically more efficient, even if it means I have to do so myself.
For me the whole value of a robot vacuum is that... I don't vacuum, the robot does, and when it does so it when I'm not at home. Reduce the dust periodically by 75% so that it's neater; I'll do a deep clean every few month or so.
(maybe what you're saying is that the cost of vacuuming for you is low enough that it's similar to launching the robot vacuum, in that case I understand you !)
> the whole value of a robot vacuum is that... I don't vacuum, the robot does, and when it does so it when I'm not at home.
I agree, that's indeed the value proposition and that's in fact why I got mine... but do you really "just" do that, namely turn it on (or automatically with scheduling) or rather do you
- very it's charged (because sometimes it doesn't actually get right on its station)
- verify its empty or that its bin isn't full
- remove chairs and cables
then and only then start it?
I had just a few couple of "bad" experiences when it got stuck on something just a few centimeters of the floor or (and that one really sucks) cables entangled inside to reconsider how "worry free" letting it roam without supervision was. I don't have pets but if I did the poop smearing videos would also scare me quite a bit.
I agree, that's indeed the value proposition and that's in fact why I got mine... but do you really "just" do that, namely turn it on (or automatically with scheduling) or rather do you
- very it's charged (because sometimes it doesn't actually get right on its station)
- verify its empty or that its bin isn't full
- remove chairs and cables
then and only then start it?
I had just a few couple of "bad" experiences when it got stuck on something just a few centimeters of the floor or (and that one really sucks) cables entangled inside to reconsider how "worry free" letting it roam without supervision was. I don't have pets but if I did the poop smearing videos would also scare me quite a bit.
I see where you're coming from, because my robot tends to eat the cats toy as well as my computer charger cable. In my case the prep is minimal, as it's just putting them on the sofa. Otherwise the robot has no problem charging and returning home, (I only had issue 3 times in 5 years ?). Concerning the bin I have pets so in my mind I have to empty the bin after each run. But psychologically I see tons of hair so in a way it's satisfying to see a job well done. The pets are fortunately well behaved so poop smearing is a non-issue (some more recent robot also have object detection to avoid this kind of stuff).
In any case, the ritual is to run it every odd day or so just before leaving for work so the mental burden is low.
In any case, the ritual is to run it every odd day or so just before leaving for work so the mental burden is low.
I worked on some next next gen AR glasses, I saw and used both orion and the prototype version of the rayban meta display glasses
Orion suffered from the problem that plagued it throughout its production: shifting goals and poor product management. The display team promised and then didn't deliver, twice.
One of my friends showed me a demo in meta display glasses (the production one) that allowed you to take a youtube recipe video, feed it into the dipshit machine and it'd spit out a step by step guide for you to follow. That was really nice to use.
The demo I worked on was taking some research glasses and attaching SLAM, always on audio and basic VLM descriptions of what you are looking at and dumping it into a database. You'd be surprised at how useful that context was. at the end of the day you could say "In the meeting I had with x what were the actions" Sometimes it even worked. Because everything that you saw or did was geocoded (ie it knew which room you were in) and linked to a calendar, even though facial and voice recognition wasn't allowed, you could get something akin to a decent executive assistant.
Making it secure and private is a big fucking challenge. I put in face and screen blurring, but thats not really enough. There is OCR based on eye gaze, which is hard to censor. The problem is there is more than enough context stored to infer lots of things.
The problem is not just securing the storage, its finding a decent sharing mechanisms that allows you to perform normal actions without asking a million questions or over sharing.
Orion suffered from the problem that plagued it throughout its production: shifting goals and poor product management. The display team promised and then didn't deliver, twice.
One of my friends showed me a demo in meta display glasses (the production one) that allowed you to take a youtube recipe video, feed it into the dipshit machine and it'd spit out a step by step guide for you to follow. That was really nice to use.
The demo I worked on was taking some research glasses and attaching SLAM, always on audio and basic VLM descriptions of what you are looking at and dumping it into a database. You'd be surprised at how useful that context was. at the end of the day you could say "In the meeting I had with x what were the actions" Sometimes it even worked. Because everything that you saw or did was geocoded (ie it knew which room you were in) and linked to a calendar, even though facial and voice recognition wasn't allowed, you could get something akin to a decent executive assistant.
Making it secure and private is a big fucking challenge. I put in face and screen blurring, but thats not really enough. There is OCR based on eye gaze, which is hard to censor. The problem is there is more than enough context stored to infer lots of things.
The problem is not just securing the storage, its finding a decent sharing mechanisms that allows you to perform normal actions without asking a million questions or over sharing.
Like for example: I don't know, sometimes I'm cooking for my partner. Do they hear a one-sided conversation between me and the "Live AI" about what to do first, how to combine the ingredients? Without the staged demo's video-casting, this would have been the demo: a celebrity chef talking to himself like a lunatic, asking how to start.
That’s fine. I cook. I’ve got my AirPods in listening to music and talk to an iPad with my recipe on it to set timers and the like. And sometimes I talk to my Dutch oven because it has feelings.
Voice control makes a lot of sense in cooking because my hands are messy. I’m just not sure what the glasses are for that could be fun or helpful while cooking. The demo video is… amusing but also wrong, because you’d need to have the recipe before you setup your ingredients.
Voice control makes a lot of sense in cooking because my hands are messy. I’m just not sure what the glasses are for that could be fun or helpful while cooking. The demo video is… amusing but also wrong, because you’d need to have the recipe before you setup your ingredients.
I tried cooking with a Vision Pro. It was nice to throw up the recipe above the prep area, hovering, so I could prep and check off without messing up a book or print a copy just for single use.
I also liked throwing up a timer above each pot so it was simple to glance at the stove and see how long each pot had on it.
But the Vision Pro is bulky, hot, and heavy. I wouldn't try it again. I could imagine giving it another go with glasses, but I wouldn't be using the AI part of it.
I also liked throwing up a timer above each pot so it was simple to glance at the stove and see how long each pot had on it.
But the Vision Pro is bulky, hot, and heavy. I wouldn't try it again. I could imagine giving it another go with glasses, but I wouldn't be using the AI part of it.
The glasses could present a recipe as actions needing to be taken in the field of view, instead of as a bulletpoint list in a book. And the glasses could use the video stream from the glasses to improve help an AI may give during cooking.
Then again, I don't see myself using them. I also cook and I'd rather just internalize processes and recipes, and something like this would make it way too easy to just rely on the glasses to "know" everything.
Then again, I don't see myself using them. I also cook and I'd rather just internalize processes and recipes, and something like this would make it way too easy to just rely on the glasses to "know" everything.
> My outsider view is that there are more people and more volume of crypto involved in the speculative & scam sector than in the human-rights sector, but I'm willing to be wrong.
I would assume this is true, but is it the right metric?
Would it be ok if it was 51:49? If $1m of crypto lets 100 Russian dissidents get out of Russia does it matter that there are $10b of pump and dump schemes? What if it's only $10m? What if its grannies having their life savings scalped? This feels like trolley problems all the way down.
I would assume this is true, but is it the right metric?
Would it be ok if it was 51:49? If $1m of crypto lets 100 Russian dissidents get out of Russia does it matter that there are $10b of pump and dump schemes? What if it's only $10m? What if its grannies having their life savings scalped? This feels like trolley problems all the way down.
I think it's pretty obvious that something like this product is going to be the future, but the technology is still pretty raw. Eventually, humans will have some kind of personal assistant baked into their field of view. Maybe we're on the cusp, or maybe we're 50 years out. Hard to know.
I thought this article would be about preventing your glasses getting steamed up when cooking
I think there is a big gap between :
- how messy the World is (imagine a "normal" kitchen, not a prop for a demo)
- how good existing low-tech solutions are (e.g. just a book or a printed page for a recipe)
meaning that to provide marginally more value the tech has to be excellent at handle a messy World. Sure the tech is radically improving but we are vastly underestimating, as we did for robotics for decades now, how challenging seemingly menial task can be and how to do so reliably.
I always think of my Roomba in this kind of situation, namely when one tries to imagine a future where tech works outside of the computer. Basically my little vacuum robot mostly sits idling in the corner because... a basic broom or a normal vacuum cleaner is radically more efficient, even if it means I have to do so myself.
So... yes I'm techno-optimist but the timescale is IMHO not as close as we hope for.
PS: AR setups (glasses or not) do NOT have to be connected to be useful. Do not get footed by Meta slick ads to imply that only they can make glasses, heck I'm taped a pair together 5 years ago with a RasberryPi Zero. It wasn't slick https://twitter-archive.benetou.fr/utopiah/status/1449023602... but it worked.