Liberty

Good thread by Sinofvsky about software quality, scale, etc: https://twitter.com/stevesi/status/963142502604779520

I hope you can make it Liberty! Last year was great fun. Having a group of people who all like the same stock is a very powerful filter for finding like-minded investors. Unfortunately, the CSU AGM is at the same time as the FFH meeting. And the evening before will compete directly with the Premier Fairfax Financial Shareholder’s Dinner. Still, I'd like to do the CSU get-together even if Liberty can't make it. I know this would be even trickier for you but if you can make it for the CSU AGM, you might consider coming two nights earlier. The Ben Graham Dinner on April 24th is a great event too. The reason why I'm not sure if I'm coming is because my wife is expecting our second boy in March. If I can get away at all, I'll stay just one night. It's too bad about the two events being at the same time... If I arrive early enough in town, I might drop by NoRm's pre-dinner meeting at the Fairmount. Then later in the evening I'll probably go have a drink, probably again at the C'Est What, and anyone from here or FinTwit who wants to join is very welcome. It was a pleasure to see you last year, I hope to be able to make it this year again. I'll try not to drop my scotch all over myself this time. Cheers!

The link is now open to non-subscribers: https://www.scuttleblurb.com/trup/

I think a lot of parts of the healthcare industry in the US are over-earning. The parts that are making things more efficient and keeping costs down are probably safe, but the parts that have seen rapid price inflation for decades probably will get on a new trajectory at some point.

That's not how it reads to me.

For those who, like me, are curious about the AGM this year, it's going to be on April 26th at 10 a.m. I was told: “It will start around 10am. we’ll have our six Operating Group managers available. Following general Q&A .. will organize six … break-out meetings, each hosted by one of the Operating Group managers. The content & format will be left to the discretion of the managers” I'll try to attend, but it might depend how things out of my control. If I can go, I'll do a little drinks get together the evening before like last year.

New acquisition: https://seekingalpha.com/pr/17072486-liberty-latin-america-acquire-controlling-interest-cabletica

http://www.yetanothervalueblog.com/2018/02/venturing-through-gci-proxy-lvnta.html

And here's a recent presentation by Housel, if you want more:

YCG letter, contains a nice overview of the Priceline model: http://www.ycginvestments.com/media/39-2017-12-31-Client-Letter---Final-(website).pdf

http://www.businessinsider.com/apple-outsold-the-entire-swiss-watch-industry-in-2017-2018-2

Good piece: http://www.collaborativefund.com/blog/why-competitive-advantages-die/

John McFarlane, the founder and ex-CEO of Sonos, has some first impressions on HomePod: https://twitter.com/JohnLMacFarlane/status/962077667330727936 (thread)

Couple more bits of news: https://www.reuters.com/article/us-amazon-brazil-exclusive/exclusive-amazon-eyes-new-warehouse-in-brazil-e-commerce-push-sources-idUSKBN1FT2HM https://www.reuters.com/article/us-delivery-service-amazon/amazon-tests-delivery-in-los-angeles-shipping-shares-sink-idUSKBN1FT2DI https://www.theinformation.com/articles/amazon-is-becoming-an-ai-chip-maker-speeding-alexa-responses https://www.wsj.com/articles/amazons-latest-ambition-to-be-a-major-hospital-supplier-1518517802

https://www.cnbc.com/2018/02/09/uber-waymo-lawsuit-settlement.html "Uber will pay Waymo a 0.34 percent equity stake, the companies said on Friday,"

Sounds like you have the gambling gene. When you notice that, the proper thing to do is to stay away from gambling.

I'm not saying Homepod will be great or not, I'm saying your previous objections are weak. Apple is trying to make a speaker that sounds good (at the higher end of this category) by using computational acoustics not used by anyone else, and it happens to have voice assistant capabilities, and maybe it'll grow on that side over time. The others are trying to make assistant-speakers that can produce ok-ish sound quality. Not the same target and strategy. Not saying one's better than the other, but Apple certainly knows that music is important to its users and with the AirPods and HomePods it's giving them something useful, even if not flashy. And Apple has never been early to categories. iPod wasn't first MP3 player, iPhone wasn't first smartphone, iPad wasn't first tablet, Apple Watch wasn't first smartwatch, etc. They're particularly late on this one, but I suspect that they're stretched pretty thin these days working on a new iPhone/iPad/Macs/Watch every year + AR stuff + VR stuff + car stuff + Airpod stuff (small item, but tons of engineering in there) + Homepod + Apple TV + new OSes + etc. Can't do everything, no company can. Exactly. A speaker doesn't have to run ML algorithms on photos and videos, which uses mostly GPU. A speaker isn't storing data, it's a way to play it. If data was going to be stored related to a Homepod, it would be in the cloud, as I doubt there's much local storage, and most of it will be music cache that gets frequently overwritten. Alexa has a very different approach. It's one language, it's almost fixed syntax (you have to learn a vocal command line). Siri is multi-language and can accept much fuzzier syntax to try to deduce intent. A lot of people compare Siri on a phone to Alexa on a plugged in speaker with 7 microphones and beam-forming... I suspect Siri might get better if they decide to focus more on it, but it's definitely not been a company priority and they squandered their early lead (when it came out in 2011, nobody had anything like it on that scale), so there are definitely pluses and minuses in their column, but the analysis certainly isn't as superficial as you make it out to be. None of the voice assistants are that great all around right now.

That's a great view. Direct youtube link: https://youtu.be/Z_kfM-BmVzQ?t=13s

Apple was fine partnering with Google up until the point when Eric Schmidt, from his seat on Apple's board, decided to compete with Apple. At that point it was pretty hard for Apple to let their main competitor control important software on their platform, so they had to build their own. I don't think they necessarily wanted to do that, but it was a strategic imperative.

Although PR is part of the US, it's over 1000 miles away from Florida and isn't a state. Nearly half of all Americans aren't aware that Puerto Ricans are US citizens. Similarly, Puerto Ricans look different and speak a different first language than do most Americans who live in the lower 48. See the Jones Act as an example of the US government being generally indifferent to the interests of PR. https://www.nytimes.com/2017/09/26/upshot/nearly-half-of-americans-dont-know-people-in-puerto-ricoans-are-fellow-citizens.html I know all that. Still doesn't seem like a good reason. But then, the same thing happened with New Orleans..

I don't understand why the US seems to have forgotten about Puerto Rico... Listen to this: https://overcast.fm/+LHycafZNs

New post about installing solar panels on his roof: http://www.mrmoneymustache.com/2018/02/07/diy-solar-power/

I think you under-estimate what the set of potential "tasks at hand" will be for an audio-interface computing device. You seem to assume that the extreme thin client model is the obvious end game here. I'm not sure that's true, because 1. Apple has gone out of their way to state their philosophical/privacy objections to the thin client model and 2. Apple has demonstrated comic ineptitude in most of the domains that model requires to work. This can't be emphasized enough because I still hear Apple people in denial about it: Siri is terrible. It is really really really bad. Frankly, they should be thinking of all of the on-device stuff they can be doing to differentiate, not deciding to fight the battle on the terms that give the most advantage to their two competitors. In general, developers figure out ways to use computing resouces as they come available. Nobody could have told you in 2007 -why- Excel in 2016 was going to require 8 times the amount of RAM to run. But anybody familiar with the history of software could have probably made such a prediction correctly. Likewise, I can't tell you, with certainty, what the TalkyTube products are going to be doing in five years. I can just tell you that I suspect HomePod1 will not be able to hang, and maybe that wouldn't have been true if they'd just given up an extra $10 of unit profit. That's a shame, because this looks to me like yet another space that was a totally easy lay-up for Apple that they're inexplicably botching. (TV still being the champion) You don't know how products are iterated. It's also a shame the iPhone didn't come out with a GPS, the ability to shoot video, a hiDPI screen, an app store, or a more powerful SoC, right? Watches are also kept for a very long time, yet the first version of the Apple Watch wasn't the end-all-be-all, right? But they're growing over 50%/year and are probably going to be the most profitable Watch brand in the world soon (if not already) within a few years. Maybe Apple will miss with the Homepod, but it won't be because version 1 wasn't perfect and didn't have enough FLOPS of computing power. Putting an A11 in the Homepod would just have made it more expensive and there would be nothing to use the extra power, it would have just laid there wasted. Software doesn't magically appear, and if you think privacy requires local instances of huge power-hungry software, you don't understand that either. I'm pretty sure the Homepod is massively more powerful than any smartspeaker by Amazon, Google, Sonos, or anyone else out there. Probably by orders of magnitude. Apple Maps and Siri are inferior to equivalent Google offerings because Google is the best at those, had many years of head-start, and because each company's DNA is set up to be exceptional at different thing. The same reason why Android is inferior to iOS and Google's hardware is inferior to Apple's hardware. If you're looking for a perfect company that is better than everyone else at absolutely everything, let me know when you find it.

The A8 is very powerful for audio, especially since most of the other stuff will happen cloud-side (search queries and AI-related stuff) anyway and there's no screen with millions of pixels to refresh 60 times a second or whatever. I think you under-estimate how powerful SoCs like the A8 are for the task at hand. Especially since an A8 on battery gets heavily power-managed and throttled while an A8 plugged in the wall will be able to go flat out with no thermal worries... If anything, it's overkill.