Their support for advertisers is one of the worst in the business.
6,234 karma · joined July 23, 2016
It’s whether you absolutely need the latest capabilities (at the cost of very high prices, sending your data to them, and being totally at the whim of 2 companies, that can shut you off anytime for any reason).
With how good LLMs are already, there’s tons of tasks where not being at the absolute bleeding edge doesn’t matter, especially when you add cost/freedom/supply chain risk/not leaking your data.
Even more - there’s increasing number of companies that give you ability to post train open weight model yourself, for your own use case. Given how many of the gains today are from post training, if you post train it for your specific use case, you’re very likely get model that you own, that works for you as good as frontier, at the fraction of the cost.
That’s not something for an average Joe to do, but for any bigger business with big spent it’s only natural thing to look into. Just one example - cursor composer - that’s fine tuned kimi.
It’s not whether frontier labs will stop releasing models. It’s whether they can generate enough profit out of them. 2 years ago (even 1) they basically had monopoly and combined with demand explosion as capabilities exploded - valuations grew to insane levels. But math now looks different - they no longer have monopoly.
10 years ago it was super hard to have usable “frontier” ML. You needed very complex data warehouse, feature engineers, feature stores, multi level ranking, calibrations, tons of different model architectures, etc, etc. Each by itself was extremely hard engineering problem and really only handful of companies could deal with that complexity.
With LLMs, 95% of that is gone, infra to support them is greatly simplified. Of course, to make really reliable, performant, user friendly, etc - you still need to a lot of engineering. But it’s very different challenge.
LLMs are still surprisingly “easy”. You need maybe a couple dozens of right people, a lot of good quality data and a lot of GPU that you know how to operate. There’s relatively little “secret sauce” needed.
You cannot just put more satellites on top of cities, as they aren’t geostationary (which is why they’re fast), so you need to add insane amount of satellites, that would do nothing most of the time, when they’re not on top of densely populated area.
On top of that, their claims it’s profitable does some heavy lifting with how to account for the cost of launching rockets.
It’s extremely cool product and very useful for many customers. But sustainability of current pricing is very questionable.
There’s nothing fundamentally wrong with distillation.
- to take as my personal device when I travel
- do personal stuff when my corporate laptop is connected to my home setup and constant switching between computers is a hassle (even with a built in kvm in a monitor). i know, 1st world problems
It has a price point that makes it no brainer for me.
1. She stopped making money for rich people.
2. She herself wasn’t rich enough.
Leon is too rich, and he keeps on making money for the right people.
But let's go back to the original point, about being able to UPGRADE (not repair/replace) battery in the car. 20 years old car is worth like $1k-2k, which is fraction of the cost of the new battery.
While it's cool thing to do for hobbyists, it makes 0 economical sense.
Focusing on being able to upgrade battery (and to be clear - upgrade, not replaced/repair) is solving 1% problem.
That's how things worked pre-AI, and old problems are new problems again.
When you run any bigger project, you have senior folks who tackle hardest parts of it, experienced folks who can churn out massive amounts of code, junior folks who target smaller/simpler/better scoped problems, etc.
We don't default to tell the most senior engineer "you solve all of those problems". But they're often involved in evaluation/scoping down/breakdown of problem/supervising/correcting/etc.
There's tons of analogies and decades of industry experience to apply here.
That being said - feature I LOVE added recently-ish that made really happy I’ve upgraded my many years old garmin was a flashlight (proper one, not screen brightness). It seemed like a gimmick but it’s now one of most used features on my watch - walking dog at night, looking for kids toys under the bed, fixing things around the house, looking for things in the bag, etc.
And yes, I do legalese details of that are much more complex. But it just makes no common sense.
It's nearly impossible to bankrupt huge country like Russia. Unless there's civil unrest (or west grows balls to throw enough of resources to move the needle), they can continue the war for decades.
What Russia is doing is each week borrowing more and more from the future and screwing up next generations on a huge scale by destroying it's non-military industrial base, isolating economy from the world and killing hundreds of thousands of young man who could've spent decades contributing to the economy/demographics.
Where they suck is high level problems like - is the code actually solving the business problem? Is it using right dependencies? Does it fit into broader design?
Which is expected for me and great help. I'm more happy as a human to spend less time checking if you're managing lifecycle of the pointer correctly and focus on ensuring that code is there to do what it needs to do.
That’s the thing about any automations that are just aides. Humans are extremely bad at monitoring machines, and if aide system is good enough that trick you into thinking it’s actually stand alone and in control, you get complacent very fast, stop pay attention as you convince yourself that automation got it.
So bad level 2 driver assists are so bad, that no one will get complacent, as they give you only very minor help. Really good ones (like comma) can trick you into thinking that they can do much more than they’re designed to do.