lkois

What would be an unbenchmaxxed model here? A general image classifier like Inception v3, without the fine-tuning?

The 200 test images here are 'wild'. Fine-tuning in-domain is the whole point. This is what model training is. A skin cancer diagnosis tool doesn't also need to generalise to dog breeds to be valid, and will almost always perform worse if it does.

Not sure if this is connected, but I've found Gemini pretty hopeless within its own notebooks. I've been doing some job applications, and added cv and docs into a notebook. I then create a new chat to say "here is a job description, help me write a cover letter" or some such.

After about 3 messages in any given chat, a follow-up to "rewrite that with a more friendly tone" will result in a letter for a completely different job from another chat within the notebook.

No they don't. The two Huawei models mentioned are completely different form factors, that are much more open-ear and further from squishy silicone than any Airpods.

This is exactly the assumption that the carefully worded juice hopes you'll make. That Apple's claim to innovation precludes even a cursory glance at the competition

If self-hosting is the only solution you can think of to maintain model stability, it's no wonder you missed the point. We already do all those things you mentioned, even with cloud providers. Model versioning exists for exactly this purpose, just like it does with any software dependency. We can specify sonnet:v1.23 and decide if and when to upgrade.

But even if we only asked for sonnet:latest, the last thing we'd expect is opus. Model names should be indicative of breaking changes, and change management doesn't just go in the bin because of non-deterministic tools.

That's a narrow take. Non-deterministic doesn't mean random; workflows can be reasonably validated and consistent to some known degree.

I work for an education department that serves a chatbot for students, and model changes go through painstaking content safety reviews. I initially assumed it's just a bunch of bureaucratic paranoia. But every other model upgrade has a measurably different adherence to the existing system prompts about not talking to the kids about sex and drugs and mental health issues.

A couple of years in Tokyo gave me great trust in certain weather apps for eerily precise minute-by-minute rainfall predictions. After a few years back in Sydney, my trust in them has dropped back down to the usual uncertain baseline, and my reliance on them has been mostly replaced with my own rough assessment of the air and sky.

On recent visits to Tokyo, with my current habits, the general crowd (who presumably check some app) has been the more reliable forecast. Either everyone has an umbrella in their hand or they don't. And I discover whether I should have brought my umbrella or if I'll be lugging it around pointlessly, just far enough from my hotel to be stuck with my choice

I wasn't aware of this game until AFTER I saw the movie. Made a lot more sense after that. Watching a guy solve the same puzzle over and over felt like an odd narrative choice.

A labyrinthine Shinjuku would still be way cooler. Shibuya station doesn't have quite the same disorienting featureless repetition to it, but offers a more overwhelming MC Escher-like experience where you can often SEE exactly where you want to go in the near-distance, but every path and every turn brings you into a completely different space, and rarely any closer to your destination.

If you just want to clone a typical garage door remote, you're way better off buying a generic $5 fob. Almost all the 433MHz garage fobs I've seen are designed to easily clone each other. And if cloning doesn't work, it's probably because you need to access the door controller and hit a few buttons to pair the remote directly. Also trivial, and also often pretty physically insecure. In the smaller apartment buildings I've lived in, the boxes containing these control units have usually just been hidden behind by a couple of screws and a "danger electricity" sticker.

If your current remote has multiple unused buttons on it as they often do, you can already try this out.

Since you cannot imagine how they'd perform, isn't this the perfect opportunity to test your assumption?

Who said that?

What do you mean by proportion? They are different data sources, and their usage is determined by system design.

eg A driving decision system needs to know object distances AND traffic light colours. It doesn't particularly need to know the source of either. You could have a camera-only system that accurately determines colour and fuzzy-determines distance. Or you could have a LIDAR-only system that accurately determines distance and fuzzy-determines colour.

Or you use both, get accurate LIDAR-distance and accurate camera-colour and skip all the fuzzy-determination steps. Or keep the fuzzy stuff and build a layer of measurement-agreement for redundancy.

So then the question becomes, what's your proportion when deciding whether to stop at a traffic light? Is it mostly light colour or mostly distance to other objects? Or 50/50?

I'd say it's 100/100.

> To the extent that "ban cars" even exists as a real political archetype rather than a meme, this is just patently not true. At least one of the two co-hosts of The War on Cars (again, a title which is intentionally tongue-in-cheek) has a preteen son.

Similar for the YouTube channel NotJustBikes, who has gone into great detail about the advantages of raising kids in a city planned around pedestrian and cyclist usage, and not in a suburban sprawl.

Sure they may become mercenaries. Maybe "Hyundai factory worker" is akin to "remote oil rig worker".

Or they could just be regular people with regular, stable jobs because they haven't chased a life of risk.

But China banning foreign apps also plays into their stranglehold on their domestic media and economy, so it's not a purely adversarial move against the US.

Or to put it another way, should the US also ban/censor Chinese art and cinema within it's borders?