> there's no good way to do LLM structured queries yet
Because LLMs are inherently designed to interface with humans through natural language. Trying to graft a machine interface on top of that is simply the wrong approach, because it is needlessly computationally inefficient, as machine-to-machine communication does not - and should not - happen through natural language.
The better question is how to design a machine interface for communicating with these models. Or maybe how to design a new class of model that is equally powerful but that is designed as machine first. That could also potentially solve a lot of the current bottlenecks with the availability of computer resources.
That's their choice, but they also choose to suffer the consequences. Expecting the world to cater to your needs specifically is such a typical boomer attitude and should no longer be tolerated.
And, expecting people who are happy with what they already have and have already paid for to switch to your newer, more complicated, more expensive system so that your numbers go up is another attitude that should not be tolerated.
Horses, no. That would impose quite a lot on everyone else. But walking, or taking the bus, vs. owning an expensive personal transportation device... yes.
While we're at it, let's get rid of the ADA. Those disabled people expecting the world to cater to their needs specifically are so abusive to those of us with perfectly functional bodies and flexible minds.
Using a battery powered electronic device as a “pass” detected by another handheld electronic device, both of which are contacting cell towers, exchanging data with data centres 100s of kms away, filling out detailed profiles of user behavior … rather than a paper ticket?
You will be the "boomer" some day. I wish people had more empathy.
An example: Presbyopia came on hard for me in the last couple of years Now I really appreciate low-vision affordances that, as a younger person, I couldn't have cared less about and would have seen as an unnecessary cost.
I used to laugh about the 'picture signs'; like the universal nose in book sign that means library. Or the airport logo on the exit sign on the freeway.
Until I spent some time in a country whose predominate language (and signage) was not english.
Maybe those pictorial signs are a good idea after all.
When OP is 85, I hope some whippersnapper 20 year old says to him, "Come on, grandpa. You need to get that neural advertisement brain implant like the rest of us, or you can't buy anything. Why should businesses need to support your lame smartphone? Step into the 22nd century, pops!"
It is simply false that it was Merkel who decided to shut down nuclear power plants. The decision had been made over a decade earlier. She just accelerated the plan in the end after a previous unsuccessful attempt at rolling back part of it. It also wasn't even really her decision, it was the will of the people that sharply turned against nuclear after Fukushima, she just implemented it.
They wasted their first mover advantage by focussing on what amounts to building toys for consumers like Sora instead of actually useful products that go beyond simple chat bots.
I think they are in serious trouble, especially with the size of their cash burn. Their planned IPO could easily turn out to be their WeWork moment where the bottom suddenly falls out on the valuation if they cannot make their operation look more like a real business before investors lose confidence.
The EU is really more middle-of-the-road in most things, while the US tends to be more extreme: more really good ideas, but also more really bad ideas. But that is also the result of the EU being largely controlled by bureaucrats and compliance officers instead of real leaders.
Performance is really not Java's issue. Even bad Java code is still substantially faster than the bulk of modern software that is based on technologies like Python or JavaScript/Node.js.
This might also be why I heard colleagues saying “Nono, listen, these ‘N+1 problems’ and our nested service calls aren’t an issue because it works well enough” until it eventually didn’t. I’d rather not have bad code in any language.
They are not trying to buy developer goodwill, they are trying to catch up with Antrophic in terms of getting those B2B contracts, which is currently the most realistic path towards not running out of money.
1. The Register reports OpenAI is well ahead of Anthropic in B2B contracts. It's Anthropic playing catch-up, not OpenAI.
2. In any case, the announcement strongly suggests that customer acquisition had little to do with this. The stated purpose of the acquisition, as I read it, is an acquisition (plus acquihire?) to bolster their Codex product.
3. But if they were hoping for some developer goodwill as a secondary effect... well, see my note above.
Measurable responses to the environment lag, Moore's law has been slowing down (e: and demand has been speeding up, a lot).
From just a sustainability point, I really hope that the parent post's quote is true, because otherwise I've personally seen LLMs used over and over to complete the same task that it could have been used for once to generate a script, and I'd really like to be able to still afford to own my own hardware at home.
I'm using local models on a 6 year old AMD GPU that would have felt like a technology indistinguishable from magic 10 years ago. I ask it for crc32 in C and it gives me an answer. I ask it to play a game with me. It does. If I'm an isolated human this is like a magic talking box. But it's not magic. It doesn't use more energy than playing a video game either.
Thanks! I've been playing with some of the qwen models via openrouter as well.. I'll have to give 9b a go at some point, I've been mostly playing with 27b and coder-next up till now.
> Very pleased to see such performance improvements in the era of Electron shit and general contempt for users' computers.
Luckily, we have invented a completely new nightmare in the form of trying to graft machine-usable interfaces on top of AI models that were specifically designed to be used by humans.
Yet, they are still around, they are still deeply embedded in most businesses, and no matter how much they screw up, it just keeps going.
reply