I don't think that's the only reason but you're spot on about OpenAI marketing being absolutely terrible. The primary product names of "Claude" vs "ChatGPT" highlights this remarkable difference. To the point where I'm seeing Claude completely take over the generic term for agent.
I do think OpenAI is doomed due to bad leadership. What you said (that the marketing is relatively terrible) and what others are saying here (that the product is worse) is damning isn't it? Are they really failing on all fronts?
The marketing of Claude relies primarily on fear, and I don't think that will have lasting success. Using fear like that tends to backfire once people see past false taking points.
The smart home is a thousand small problems to solve and should never be one catch all.
The automatic cat feeder works well. So does the roomba. I like my automated blinds but will stick with manual light switches. I consolidated my home theatre remotes. Note how they’re all seperate problems.
The smart home is here. It’s just that it was never a use case for a singular smart home platform. It was always 1000 seperate problems to solve that in no way ever belonged together and the experience was always worse when trying to combine it.
You can have manual light switches and home automation. I have exactly that, but given I built a new house with KNX integration. All light switches are physical switches and can on/off and dim the lights. The only difference is, they are not hardwired to the lamps, instead they go to the KNX machine. So if I want I can always decide what switch does what.
I have a singular smart home platform with one exception: my Dreame cleaning robot. Everything else is either integrated into KNX (lights, appliances, pool, window blinds, etc.) or Ubiquiti (cameras and doors).
But again, new building and I made the mission clear: everything has to fit into this system. Was an interesting work for the electrician ;).
They will possibly all converge when they expose a "tool interface" to some kind of model-on-prem(ish) device that you install in your home. Think of an OpenAI or Anthropic or Apple or Samsung-branded "cortex" or "brain" that controls everything to some degree using fast local models, but outsource more complex orchestration up to the cloud. Smart home products will integrate with these devices because its going to open up a whole new generation of the same devices they sell, just with AI model integration this time.
These devices already have a precedent, your apple tv or google/amazon speaker thing. I think we will see these probably become LLM/model/AI gateways in the future.
We are basically already there, with HomeKit plus open bridge that can make any device visible on HomeKit (like nest cameras) and usable in automations (it works the other way also, it’s just a good way to get compatibility).
I would like to explore some open source solutions though, it would mean setting up a local system on a Mac mini for speech recognition and local processing. Bonus is you could use Anthony Daniels (KITT) as your assistant voice.
Synchronizing lights is hard, so in the living room kitchen where we have 4 different lights, having a way to turn them on and off is useful. We don’t have automatic lights anywhere else, but I notice at least, where we don’t, the light switches don’t get dirty (and you don’t get dirty from using the light switch), so maybe we will make them all automatic eventually just for hygiene reasons.
>I use Home Assistant (https://www.home-assistant.io/). It has an integration for lots of manufacturers. Single app, single entry point, cross manufacturer automations.
It’s fab capacity. Fwiw dram is different enough that fabs are not transferable between dram memory and other usages. It’s nice to think ‘wow if they made the current 10nm dram on the latest 2nm processes it’d be much faster’ but it doesn’t work that way. The specific size is needed for the capacitance. Sram can be made on fabs that make other circuitry since it’s transistor not capacitor based but is less dense.
I asked for evidence different people keep feeding me opposite stories: one insists its not fab capacity but wafer competition, with a recent article claiming HBM3E takes 3 times as much wafer area per bit than LPDDR5X. Others tell me the complete opposite: its fab capacity, not wafer shortage.
Do we have citable references to ground either set of claims?
I believe those are two ways of describing the same thing. If you're able to book some fab capacity, that means you get to decide what the fab does with the next wafers in the queue.
From your sibling comment, I think you're interpreting the 3x HBM stat as contributing to making wafers scarce. It's more that the next wafer to be processed in a fab is especially precious, making the opportunity cost larger. The beach sand remains plentiful.
Fwiw I’ve seen confused or misleading posters reply to this change with statements along the lines of ‘this only means that they don’t allow tourist to apply for greencards in the country’.
Which is nonsense, it applies to all non immigrant visas such as work visas. But it’s a line you’ll see various people try to claim as if this isn’t devastating to every spouse of a us citizen who now can’t get a greencard without leaving their us based job and family.
I’ll give a different opinion that it’s really heavyweight to install exodos locally just to get a nostalgia hit when there’s plenty of sites like the above where it’s one click to run an old game fullscreen in a browser window.
True, but when I installed ExoDOS I choose the option where it just downloads the descriptions, tagged metadata and a few screenshots per game with a searchable menu system. You can browse by name, genre, publisher, resolution etc, pick the game you want and it gets only the necessary files from the torrent (which, given the era, are very small). It's quick and seamless.
I keep minimal ExoDOS, MAME and RetroArch installs on my laptop so when I'm reading a retro article about some cool game I've never heard of (or only vaguely remember) it's easy to download the game files and give it a go. Frankly, retro emulation has gotten so incredible lately with upscaling, 4K texture packs, mods, decomps/recomps and fan translations of Japan-only titles - it's been 18 months since I played a game released in the last decade. Currently, I'm halfway through the best late-90s Japanese shmups. Next up I have 126 PS3/X360 titles curated from top ten lists on my backlog.
You don’t necessarily know when someone will decide to do a commercial release of an old game, causing it to disappear from various abandonware sites. Much simpler to grab eXoDOS once and use it for life.
Same but i feel many sports are weird in that i’ve never been convinced that there’s a particularly natural right or left handedness to them.
Eg. For pool does the more dextrous hand need to push the cue or does it line up and guide the front of the cue? I can see tradeoffs each way and the front hand is certainly not just limp when playing.
Hockey is similar. The top or the bottom hand being the more dextrous probably has tradeoffs but I don’t see either grip as being more or less natural for handedness. I don’t play hockey but play golf and cricket which have similar grips and am similar there to you too.
Golf and baseball batting have obvious handedness - the muscles that pull your towards your centerline and then across your body and significantly stronger than the ones that push your arm back out away from your body, and the right-handed stance in these two sports uses the stronger muscles in the right arm.
phil mickelson, easily the most famous left handed golfer, is right-handed but plays lefty because he would stand across from his dad and mirror his swing as a kid
fun fact: vs the US, golf stores in Canada carry more left-handed clubs because a right handed hockey player has their right hand higher on the stick which is the same orientation as the grip for left-handed golfers.
Yes but crude is not really fungible. About 14% of US crude imports are effected which is about 8% of overall crude refining.
By the US not being reliant on imports I was saying that even with just local crude oil production the US can satisfy internal demand for petroleum products.
My wider point is that of course everyone knows that that's not how the economy really works and I was replying to nradov oversimplifying by pointing out that if it was that simple US petrol prices wouldn't have gone up as much as they did. Because even though it's only a few countries with specific refineries that are actually reliant on the straight being open to get their specific required flavour of crude it's everyone in the refined markets that are actually effected by the supply of that crude because it effects the supply those refined products.
It is yes. The anglosphere has historically been somewhat base 12 in currency, time and units of measurement.
Currency is now metric but there’s still a few base 12 things in common usage (feet and inches) in the us at least. Nobody’s gone to metric time yet and base 12 transfers smoothly to base 60 too.
Of course it's because of imperial units. TIL, thanks. But on a sidenote, I question the utility of knowing x11 and x12 when working with time. x15 could be useful, unfortunate they don't teach that (but I think most people with higher education learn it on their own).
Feet and inches long predate imperial units, and the US has never used the imperial system, btw. “Imperial” has a specific meaning and isn’t just “anything not metric”.
Anyway, base 12 is also built into most Germanic languages which have unique names for 11 and 12 (rather than something along the lines of “one-teen” and “two-teen”, which is more common in Romance languages IIRC.
Out of the most spoken romance languages, Spanish and Portuguese have distinct names up to 15, French and Italian up to 16, while Romanian does stop at 10. This suggests hexadecimal influence to me.
A lot of these are very reminiscent of asking ai to find references to a subject. You end up with very tenuous links that take multiple steps to get there and i’m never convinced that the author really did intend that meaning.
Of course this could also be traditional literary analysis. It’s hard to tell.
They are bog standard 2D flat panels with 3d scenes on them to anyone wondering how they can possibly work.
The billboard ads on buildings shown in the article are also 2d. They look really bad in person unless you close one eye and watch exactly at the intended perspective (or instagram it with a misleading “omg wow new 3d billboards” caption as many do).
To be fair, the companies that install these often choose locations where large amounts of people are exiting or encountering the billboard at the intended angle.
The trick also works much better if it's a surprise (your brain doesn't have time to overanalyze) and in a position that is at a sharp angle.
So, exiting a train station, for example.
The animators and designers also perform a lot of tricks (such as parallax tricks and cinematic framing) to really sell it. For example https://www.youtube.com/shorts/-YKfFEL1vjg
I do think OpenAI is doomed due to bad leadership. What you said (that the marketing is relatively terrible) and what others are saying here (that the product is worse) is damning isn't it? Are they really failing on all fronts?
reply