On the one hand, organizations are without question using LLM's well beyond what is actually necessary, and as reality kicks in they're forced to scale back accordingly. However at the same time, on intervals counted in months, we're seeing breakthroughs both in hardware and software that dramatically reduce the cost of inference.
Between corporate FOMO and the rapidly decreasing costs of actually running LLM's I'm interested to see at which side of the spectrum these two meet
The title seems misleading, and reading the article explains the reason more clearly. There's nonsense OKR's and objectives at these companies to burn as many tokens as possible. It turns out that when you make a metric out of token usage, it unsurprisingly ends up becoming extremely expensive.
Inference is affordable, and you don't need a SOTA proprietary model to get a lot of use out of this technology. While you likely will still need a human engineer for quite a while longer, I don't agree that some number of humans + an LLM is going to be (or will ever remain) more expensive than just hiring more humans.
They may as well have just said:
Company institutes an OKR that the IT division must spend over $1000/day/developer (fictious number).
Company is surprised when IT division is costing far more than it did before. Company increases this to $1500/day/developer to build a system to identify why this has happened.
I feel like vibe coding is less of an issue than vibe leadership at this point, and vibe leadership has nothing inherently to do with AI. These people are getting a vague feeling in their giblets, and then chasing it to the illogical conclusion no matter the cost or outcome.
I'm not sure that vibe leadership is a new thing and in fact may be a redundant term. I've worked for enough companies to get the sense that doggedly following vague feelings in their giblets is what leadership has been since 2008.
I won't deny that sometimes it works but there's much more coverage on when it does that when it fails which only serves to amplify the survivorship bias around it.
And in a way this feels like a good thing (from a corporate strategy perspective). If MS really wants to compete with Claude Code they will need to dogfood to have even a hope of ever catching up.
As much as I may dislike MS, their software or their practices I have to admit that they have pulled this off at least once before. Back in 2019/2020 their Teams web client was absolutely atrocious and utterly unusable on Linux. Sometime in 2023/2024 it had become quite tolerable and worked mostly better than Google Meet. (Screen sharing options in Teams suck to this day, though.)
But aren’t the revenue numbers that have investors foaming at the teeth based on that “tokens as a metric” world? It can’t be both an explosive growth business and also only ROI with more disciplined spend.
The media seems hellbent on torching AI. My news feeds are nothing but stories about the evils of data centers, how useless AI is, and how much everyone hates it.
The media is hellbent on torching it, and on propping it up against all reason too, both things can be true. HN is no exception. It's another noisy room problem where the distortion in dialogue is rapidly leading us into a distorted reality. https://thenoisyroom.com/
For people who are actually interested in reality, participation in the mainstream discourse either way is a strategic error. The best thing to do is to check out from all of it, actually read the literature and listen to the technical heros who are working at the edge, and stop reading the pro/anti marketing noise from the media or corporate PR
> and listen to the technical heros who are working at the edge
that's terrible advice. those guys dedicate their lives to the advancement of this field. there's no way you will get a tempered, balanced answer from them. none of them will gravitate towards "yeah, maybe we should stop or slow down for a while".
Again you're probably thinking of the forum discussion / booster blogs / executive interviews that I'm suggesting you should leave behind. Papers are the only place left where nuance is even allowed and might actually be encouraged. Just try it, you'll be surprised. Depends on the research but.. a lot of it is kind of incentivized to align with AI skeptics actually because it leaves many things open for invention, study, and fixes
From AI companies’ perspective, it’s free press… why would they even think about stopping people talking about it!
This about it like this - if you were a CEO of a company that ONLY made garden gnomes, would you rather a) nobody ever talk about garden gnomes, or b) garden gnomes be in the news every day, people protesting because they’re losing their jobs because of garden gnomes, companies making billions and collectively investing trillions to making garden gnomes, people starting startups to support the garden gnomes pipeline, consumer electronics prices having huge variance because of the demand to support garden gnomes etc.
When you’re one of the largest garden gnomes companies in the world, you want garden gnomes to saturate the zeitgeist
Seems like a strategy that could backfire, if Congress passes legislation outlawing the manufacture, sale, distribution, possession, and admiration for garden gnomes. PT Barnum only thought there was not such thing as bad publicity because he was pulling up the stakes and leaving town before anyone woke up.
It drops compared to the immediate vicinity, but mountain areas in the hottest parts of the US can still be very hot, and even non-mountain areas in the coldest parts of the US can still be very cold, so perhaps we should just say the coldest areas.
I am afraid that the TL would be uncomfortable if they have no human team members but only agents, which means they have no space to pass the bulk and have to take responsibilities for the business results.
Solving one of the most famous Erdos problems that has remained unsolved for 80 years without using tools like lean but instead a giant reasoning block is quite a lot more than "kinda nothing"
Maybe I'm misreading but that is an absurd ToS in this context. So they're telling us they have a solution to a problem, but don't trust it enough to solve it? I tend to be averse to analogies but this feels like hiring an engineering team to build a bridge, and they tell you they're not liable if the bridge fails and collapses when used to spec.
If you don't actually believe in your product's capabilities, why sell it?
Calling the technology "text auto complete" is not productive to the discussion. Less than a decade ago the idea that a computer could take a fuzzy human-readable description and turn it into executable code was science fiction, but now it's common place. As is the ability to write long form text, and be so hard to distinguish from real that placing an em dash in your text will cause an uproar on this forum. You can describe things by their fundamental functions and make many things sound elementary but I find it counter productive given the capabilities we've seen from this technology
> Calling the technology "text auto complete" is not productive to the discussion.
If pointing out the flawed approach to making something more productive isn't productive, then what do you consider to be productive?
> Less than a decade ago the idea that a computer could take a fuzzy human-readable description and turn it into executable code was science fiction
Cobol was sold to people on the idea that anyone could create something with fuzzy human readable description that would result in executable code. That was back in the 60s.
What lessons did we learn?
1) Leaving things to the people who make fuzzy human readable descriptions turns out to be a terrible way to have things implemented.
2) Slowly and deliberately thinking things through before, during, and after implementation always leads to better results.
It's a lesson that keeps needing to be re-learned by people who don't/can't look at things through a historical lens.
It was the same with cobol, as it was with programming in spreadsheets in the 80s, as it was with the nocode movement in the 00s, as it is now again with LLMs in the 20s, and it will be again with a future generation in the 40s.
---
> As is the ability to write long form text, and be so hard to distinguish from real that placing an em dash in your text will cause an uproar on this forum.
Long form text generation that is hard to distinguish from human authored text also goes back to the 60s.
That's when we got the first instances of the Eliza effect.
> You can describe things by their fundamental functions and make many things sound elementary but I find it counter productive given the capabilities we've seen from this technology
These people will never admit they actually don't understand technology and just attack others. It's a great strategy if you're trying to proselytize a business into a religious movement, and it seems to be working on those that induce psychosis to themselves.
"Pertains" is doing a lot of work in your argument, and you're using it wrong. The data about who viewed your profile pertains to you from the moment the visit happens. That's what that word means, so your first statement is false.
The other important detail is that LinkedIn already has processed this data that definitely pertains to you, whether you paid for it or not, and are trying to sell it to you. In fact, to quote the article, LinkedIn's argument for not giving it to the user is "on the grounds that protecting that data took precedence". LinkedIn isn't withholding viewer data to protect viewer privacy. We know this because they sell it. If the viewer's privacy interest were so compelling that it overrides your Article 15 right (which is what Noyb is referring to), it would also be compelling enough to prevent LinkedIn from selling that same data to Premium subscribers.
The argument being made for this specific feature (not the ones you added) is that you can't simultaneously claim the data is too privacy-sensitive to disclose under GDPR and then sell it as a product feature
Where did you read this? From what I read in the paper it appears to explicitly state that they used NVIDIA GPU's and their MegaMOE code, which is written in CUDA.
Claude's diagramming tool that they have built into their web UI is my goto for this task. It's reliable enough that I often will delegate to it first with what I need written in prose instead of using mermaid/lucid diagram
The strait of hormuz is still closed, and a new government has not been installed.
From a conventional perspective Iran is by all means "losing" the war. However, the United States and the majority of the world desperately want the strait to be opened and have so far been unsuccessful in preventing Iran from blocking it. The US is also greatly interested in regime change, which has also been unsuccessful.
> The US is also greatly interested in regime change
Trump doesn't car about regime change. Just like in Venezuela, his plan is to kill leaders until there's one that can make a "deal" (whatever it means).
Between corporate FOMO and the rapidly decreasing costs of actually running LLM's I'm interested to see at which side of the spectrum these two meet