You must admit the way GP framed it strongly implies Meta gave the Indian government carte blanche access to intercept decrypted messages. That is a massive, order-of-magnitude different story than the Indian Gov't hacking phones (installing spyware, etc.) to exfiltrate messages decrypted on device. They are very different stories with very different implications.
Totally agree, I mentioned it in another comment but Gemini was a game changer for allowing me to increase the size of the project I can feasibly have AI work on.
Only issue is Gemini's context window (I've seen my experience corroborated here on HN a couple times) isn't consistent. Maybe if 900k tokens are all of unique information, then it will be useful to 1 million, but I find if my prompt has 150k tokens of context or 50k, after 200k in the total context window response coherence and focus goes out the window.
Godspeed. I find it frustrating these systems don't follow some sort of international standard, still. I can't justify replacing the system in my 2016 Highlander just to get Android auto, but the Bluetooth system handles non-phone calls poorly and the album art database hasn't been updated for basically the entire life of the vehicle. I don't like having to rely on the manufacturer nor Google/Apple for connectivity basics.
> I find it frustrating these systems don't follow some sort of international standard
They do, actually!
Most car manufacturers, including Toyota, now use the MOST-bus for infotainment system interconnects, which is an open standard. So your car being a >2010 Toyota, it will have a MOST system.
> I can't justify replacing the system in my 2016 Highlander just to get Android auto
You don't have to, there are plenty of aftermarket modules available that will add Android Auto to your existing MOST infotainment system. Can be a hit-or-miss though in terms of quality and integration. If you want something reliable and well integrated, stick with the OEM solutions.
> I don't like having to rely on the manufacturer nor Google/Apple for connectivity basics.
You're already relying on the car OEM for mobility and your safety, and you're already relying on Google or Apply for connectivity (aka: your phone's OS).
Remember that your 2016 Highlander (3rd gen XU50) was introduced in 2013, so developed well before that. Android auto didn't even exist back then, and we still hadn't consolidated on the iOS/Android duopoly we have today. Heck we were still using Windows phone and Blackberry OS back then. How could Toyota engineers have designed something that would still remain compatible with devices in 2025? Quite remarkable actually that it still works at all.
Is possible to equip an old car with some sort of infotainment+map system (ideally integrated with GMaps)? Something off the shelf that requires installation but after that integrates with your phone (Android or in the future Apple).
If you don't mind spending the money, go to crutchfield's website and pick a head unit with the features you want, they'll sell the kit to make it integrate with your cars built in systems.
Work wise I'm building a discrete event simulation with pysim. I manage the autonomous vehicles program and underground control center for the subarctic mine I work at, and we've desperately needed some means to evaluate efficacy of designs and forecast productivity. Using it to model scoop-truck movement and loading operations, and it's working very well. It's been a joy.
I've been taking advantage of the enormous context window of Gemini 2.5 pro to have it make upgrades, fix issues, and just generally doing all the legwork.
Personally I'm having fun learning some web design and three.js as I try to debug the very tricky issues with my new personal website loufe.ca using AI is fantastic but you need to micromanage when things get complicated, I find.
No. Five years ago BYD introduced their "blade battery", which is a lithium iron phosphate battery built up of plate-like "blades" in rectangular casings.[1] Wh/L is about the same as lithium ion, Wh/Kg is not as good, and Wh/$ is better. It will survive the "nail test" and does not not go into thermal runaway.
Today, most of BYD's products use this technology. It's been improved to handle higher charging rates. Seems to work fine. Lithium-ion has better Wh/Kg, and it's still used in some high-end cars, mostly Teslas. BYD's approach has captured the low and medium priced markets.
BYD has announced that they plan first shipments of cars with solid state batteries (higher Wh/Kg) in 2027. Price will be high at first, and they will first appear in BYD's high-end cars. Like these.[3] BYD has the Yangwang U8, a big off-road SUV comparable to the Rivian, and the Yangwang U9, a "hypercar". Just to show that they can make them, probably.
I'm so grateful to live through such exciting times. I can open HN every two to some exciting new news about ML/transformer models. I really should read more into it, but does llama.cpp use a "custom kernel" per se, with cublas, or is it just making good use of the cublas kernal?
Prescient timing, went for a 10pm workout last night and was up until 2:30am wide eyed in bed unable to sleep. I feel like I need to try this every couple years to be reminded how it affects my sleep.
I'm baffled seeing this. I'm absolutely convinced of Google value prop, knowing Gemini 2.5 pro was trained and inference is performed on their TPUs is enormous. It's far and way the best model (Claude still has my heart, but their usage limits and limited context can't compete).
The company that can allow collaboration up the value chain only like Google can in this space right now is going to win. The author went over similar pushes by Meta, AWS, and Microsoft.
Now that their eyes are on the prize, and the prize is so staggeringly big, I'm convinced of the threat to their moat.
