For all of those following, I emailed Rob and he confirmed that the focus of the project to start is reverse engineering binary blobs on existing android devices, but he is currently only at the discovery phase of picking which phones to start with. He is first checking LineageOS compatible phones using his toolset here: https://codeberg.org/rsavoye/librephone/src/branch/main/doc/index.md
Huh, they still work for all my banks here in the US. Not sure how you're supposed to access your bank account on your PC otherwise. Some banks you'll have to use the "use Desktop version of the site" option in a mobile browser to get it to work, but it will still work.
That said, didn't hird basically die because Linux gained critical mass faster and peeled off the core developers? It would be nice to imagine another bottom-up mobile OS emerging and stealing the thunder of this one, but it seems like the hope here is that Libre Phone will gravitate in some of the devs from the existing top-down open phone projects. Who knows if that will work.
One thing I wouldn't count out right now though: China is very much in favor of getting software and hardware monopolies out from the control of US companies. Free/open(ish) LLMs are the big example, maybe they will jump on this to try and break Google's stranglehold on the mobile market.
If the owners primarily want to make money by taking out a portion of revinue as dividends or distributions, like a family business typically does, then stable revenue is more important in some ways than reinvesting in growth.
If the ownership wants to make money by eventually selling their stake (shares or equity) in the company then growth is fundamental to the strategy.
The biggest difference is that the Dot Com bubble was strongly focused on tech companies going public and pumping small cap stock prices up.
The AI bubble on the other hand is almost entirely being built by private equity, with the largest players all privately held but with large cap stock companies holding substantial stakes. Rather than a bunch of small companies getting pumped up stock prices of many multiples of their debut price then falling to zero, instead we have large cap stock companies bumping up their value substantially, but not by major multiples, while the actual value of the biggest players in AI are all speculative and can't be invested in by retail investors.
This is all by design, the financiers of the AI boom are well aware that a public stock oriented rush into AI for retail investors would lead to massive speculation and an inevitable crash, instead with all the retail money going into large cap stocks they hope to capture that value and funnel the money into buying long term gains by making sure that those big companies have some stake in the "winning" private companies. When the first big AI companies go bust, they will be consolidated into their investor groups and harvested for innovation to transfer over to the winners.
Overall this strategy seems sound to avoid a major retail stock bust, but isn't wothout its own risks, for example if open source AI ends up winning out and the biggest private players fall flat they could become toxic assets and drag down the large cap stocks, and thereby the Indexes and Index funds in favor of leaner players. In the current landscape, that would mean Microsoft going down with OpenAI while Apple goes up, Apple is waiting on the sidelines with a huge cash warchest, ready to buy.
(the best) Local LLMs are FOSS though, if bias is introduced it can be detected and the user base can shift away to another version, unlike centralized cloud LLMs that are private silos.
I also don't think LLMs of any kind will fully replace search engines, but I do think they will be one of a suite of ML tools that will enable running efficient local (or distributed) indexing and search of the web.
You're joking right? "making up answers" in the case of search results just means a dead link. If you get a good link 99% of the time and don't have to use an enshitified service, that's good enough for 99% of people. Try again is the worst case scenario.
Also, you know what would make this all even worse? Laws requiring that people prove their identity in order to consume content or pull videos... just like age verification laws now being passed in several countries. What a coincidence.
Not to mention that the scraped indexes can and should be shared. Unfortunately what OP is seeing may be a move to thwart this type of brute force scraping, and might resolve as dynamically assigned domain addresses, where the URL of a set object is temporarily assigned and streamed only to a single or group of IP addresses that request it within a given timeframe before being rotated out until found in search again and then reassigned a new URL, etc. This is a frankly stupid use of resources, but can effectively be used to prevent crowdsourced indexes from proliferating, and to punish IPs or even MAC addresses or browser fingerprints associated with downloading and reuploading videos which almost certainly have stegnographic fingerprinting embedded that associate with who the video was served up to at the time it was downloaded.
OuterTune is nice, thanks for the tip. I thought for a second it was going to have my number one desired feature for a YouTube front end: playlist folders, but alas it's all just a list once again.
Is the local view missing sort by artist and album or am I missing something?
What? I pay $23 USD for YouTube Premium Family Plan, which includes ad free video and YouTube Music for 4 people. Still pretty reasonable IMO, never going back to Spotify that's for sure, I have thought about trying Qobuz for higher quality, but the price increase across my family plus the fact that I never ever want to watch ads on YouTube makes it a difficult value prop, I'd probably rather buy one album a month from BandCamp.
I looked around a bit more, seems to be regional which is growing faster. UK Meshcore is growing fast, which makes sense since core devs are there. I haven't found a good statistics source for node numbers, probably for the best given the network goals. Meshtastic of course is older so has the inertia lead, but the security and routing of Meshcore definitely is appealing to me and seems better for emergency situations where text messaging reliability is top priority.
today I learned there are a bunch of movies named Joy Ride