3 ms·

I never worked for Samsung, but I built TVs for JVC and LG, among many other brands. I don't work in consumer electronics anymore but a decade ago that was my f

by bobdvb 2y ago

I never worked for Samsung, but I built TVs for JVC and LG, among many other brands. I don't work in consumer electronics anymore but a decade ago that was my field. TVs are a wildly unprofitable business. It's astoundingly bad. You get 4-6 months to make any profit on a new model before it gets discounted so heavily by retailers that you're taking a bath on each one sold. So every dollar in the BOM (bill of materials) has to be carefully considered, and not far back the CPUs in practically every TV was single core or dual core, and still under 1GHz. Bottom of the bin ARM cores you'd think twice to fit to a cheap tablet. They sit within a custom app framework which was written before HTML5 was a standard. Or, hey want to write in an old version of .NET? Or Adobe Stagecraft, another name for Adobe Flash on TV? Apps get dropped on TVs because the app developers don't want to support ancient frameworks. It's like asking them to still support IE10. You either hold back the evolution of the app, or you declare some generation of TV now obsolete. Some developers will freeze their app, put it in maintenance mode only and concentrate on the new one, but even then that maintenance requires some effort. And the backend developers want to shutdown the API endpoints that are getting 0.1% of the traffic but costing them time and money to keep. Yes, those older TVs are literally 0.1% or less of use even on a supported app. After a decade in consumer electronics, working with some of the biggest brands in the world (my work was awarded an Emmy) I can confidently say that I never saw anyone doing what could be described as 'planned obsolescence'. The single biggest driver for a TV or other similar device being shit is cost, because >95% of customers want a cheap deal. Samsung, LG and Sony are competing with cheap white label brands where the customer doesn't care what they're buying. So the good brands have to keep their prices somewhere close to the cheap products in order to give the customers something to pick from. If a device contains cheap components, it was because someone said "If we shave $1 off here, it'll take $3 off the shelf price." I once encountered a situation where a retailer, who was buying cheap set-top boxes from China to stick a now defunct brandname on, argued to halve the size of an EEPROM. It saved them less than 5c on each box made. For long life support of the OS and frameworks, aside from the fact that the CPU and RAM are poor, Samsung, LG and Sony don't make much money from the apps. It barely pays to run the app store itself, let alone maintain upgrades to the OS for an ever increasing, aging range of products. And we as consumers have to take responsibility for the fact that we want to buy cheap, disposable electronics. We'll always look for the deal and buy it on sale. Given the choice of high quality and cheap, most people choose cheap. So they're hearing the message and delivering.

Tijdreiziger 2y ago
Yeah, but is there a way for consumers to compare the compute performance of any given TV? If OEMs differentiated their TVs based on compute performance, consumers might be able to make an informed choice. (See smartphones: consumers expect a Galaxy Sxx to have faster compute than a Galaxy Axx.) If not, consumers just see TVs with similar specs at different prices, so of course they’re going to pick the cheaper one.
- bobdvb 2y ago
  It's really hard to get these things across to consumers. This is why we ended up with phrases like "Full HD". The average consumer doesn't know what these numbers mean, people who read hackernews aren't the 99%. Phones have helped a little bit with widening the idea of newer = better, but ask the average person how many cores their phone is or how much RAM it has? They don't know. Also, it's hard to benchmark TV performance as a selling point. Perhaps sites like rtings need to have UX benchmarks as well? They could measure channel change times, app load times, etc. That might create some pressure to compete.
3np 2y ago
Thanks for sharing. Without insight beyond being a consumer, I do think there's room for disription (ideally from within the industry itself) vs 10y ago. Comparing models from 2005/2015/2025, for example. Most people literally can't tell 4k from 1080 and anything new in the HD race mostly feels like a scam. The software capabilities are all there. I think to differentiate from the no-name stuff, longevity is going to become a more significant differentiator.
- bobdvb 2y ago
  We tried to disrupt the market, back about 10 years ago. One of the significant problems is that 80% of TV SOCs are made by one company, MStar (or their subsidiary). And there's only a handful of companies who make the motherboards with those chipsets. Anyone entering the market either buys those or isn't competitive. It's hard to be competitive because everything is so concentrated and consolidated. Since ST Microelectronics and Broadcom left the TV chip market it became a much less diverse market. We were an established company who made software for STBs, we had done a ground-up build of what was probably the most capable and powerful framework for TV/DVRs. The new design was commissioned from us by a well known open source Linux distro, who then decided they didn't want to continue with the project after they realised that getting into TV OS's was hard. We then took on ownership of that project but getting investment or even commitments from buyers was impossible. The retailers and TV brands wanted to rehash the same thing over and over because that was tried and tested. It didn't matter that we made something that was provably better and used modern approaches, it wasn't worth the effort for them. If you can't order about 500,000 TVs then you're not going to get anyone to make anything custom for you these days and you'll not make a profit. -- It was a DVR/TV framework that was designed by people who had worked for big names in the TV business with a clean slate. It would handle up to 16 different broadcast networks (e.g. satellite, terrestrial, cable) and up to 255 tuners, even hot pluggable. Fast EPG processing and smart recording to either internal storage or USB storage. It was user friendly and allowed for HTML5 apps. We pushed it as much as we could but eventually on the brink of financial ruin the company was sold to someone who had no interest in what had been built. I will always feel that something great was lost.

by bobdvb 2y ago

Tijdreiziger 2y ago

Yeah, but is there a way for consumers to compare the compute performance of any given TV? If OEMs differentiated their TVs based on compute performance, consumers might be able to make an informed choice. (See smartphones: consumers expect a Galaxy Sxx to have faster compute than a Galaxy Axx.) If not, consumers just see TVs with similar specs at different prices, so of course they’re going to pick the cheaper one.

bobdvb 2y ago
It's really hard to get these things across to consumers. This is why we ended up with phrases like "Full HD". The average consumer doesn't know what these numbers mean, people who read hackernews aren't the 99%. Phones have helped a little bit with widening the idea of newer = better, but ask the average person how many cores their phone is or how much RAM it has? They don't know. Also, it's hard to benchmark TV performance as a selling point. Perhaps sites like rtings need to have UX benchmarks as well? They could measure channel change times, app load times, etc. That might create some pressure to compete.

3np 2y ago

Thanks for sharing. Without insight beyond being a consumer, I do think there's room for disription (ideally from within the industry itself) vs 10y ago. Comparing models from 2005/2015/2025, for example. Most people literally can't tell 4k from 1080 and anything new in the HD race mostly feels like a scam. The software capabilities are all there. I think to differentiate from the no-name stuff, longevity is going to become a more significant differentiator.

bobdvb 2y ago
We tried to disrupt the market, back about 10 years ago. One of the significant problems is that 80% of TV SOCs are made by one company, MStar (or their subsidiary). And there's only a handful of companies who make the motherboards with those chipsets. Anyone entering the market either buys those or isn't competitive. It's hard to be competitive because everything is so concentrated and consolidated. Since ST Microelectronics and Broadcom left the TV chip market it became a much less diverse market. We were an established company who made software for STBs, we had done a ground-up build of what was probably the most capable and powerful framework for TV/DVRs. The new design was commissioned from us by a well known open source Linux distro, who then decided they didn't want to continue with the project after they realised that getting into TV OS's was hard. We then took on ownership of that project but getting investment or even commitments from buyers was impossible. The retailers and TV brands wanted to rehash the same thing over and over because that was tried and tested. It didn't matter that we made something that was provably better and used modern approaches, it wasn't worth the effort for them. If you can't order about 500,000 TVs then you're not going to get anyone to make anything custom for you these days and you'll not make a profit. -- It was a DVR/TV framework that was designed by people who had worked for big names in the TV business with a clean slate. It would handle up to 16 different broadcast networks (e.g. satellite, terrestrial, cable) and up to 255 tuners, even hot pluggable. Fast EPG processing and smart recording to either internal storage or USB storage. It was user friendly and allowed for HTML5 apps. We pushed it as much as we could but eventually on the brink of financial ruin the company was sold to someone who had no interest in what had been built. I will always feel that something great was lost.