This writing style, which reads like an LLM wrote it, is draining to read.
I put up with it when I'm actively using an LLM tool, because I just accept that this is how these tools write, but when reading a blog I expect to read words written in one of the vast varieties of normal human-written blog formats.
Fair point. English isn’t my first language, so I run my writing through tools to polish it. I probably leaned too hard into making this one auditable. The research and conclusions are mine, but yeah, it reads more like a technical report than a blog post.
One of the things that stood out in your article was the high amount of negations or contrasts. This is one of the worst hallmarks of AI writing, because humans tend to write in a more straightforward way, without as many caveats.
My two cents: you should avoid using AI to “polish” what you wrote. At most, instruct it to correct grammar mistakes only, instead of rephrasing your words, changing your arguments, turning paragraphs into bullet lists, “improving” sentences to make them more idiomatic… Nowadays, many technical readers would rather read the occasional non-idiomatic prose than put up with the soulless writing style of AI.
As long as you are constantly diligent in staying on top of what each provider is doing that may impact you.
It reads to me an argument for self hosting, maybe a less capable model to deal with smaller compute resources, but when starting to use that model and inference software to build a benchmark you can easily rerun when you change things to observe the impact of the change.
Maybe that’s a similar level of diligence but they feel different to me.
But if I were struggling with an assignment, stuck on a problem, I would have turned to AI after a few minutes of trying to figure it out, and thus lost out on the actual character-forming and memory-building aspects of college assignments.
Which NEC commercial display do you have? How do you like it for normal TV duties? What tradeoffs were you willing to make when you bought it compared to the features which were available in consumer-focused "smart" TVs?
MultiSync P495, with the additional daughter card for the RPI CM4.
As far as regular TV duties, the CM4 running libreELEC does everything I need and then some. The HDMI-CEC was flaky at one point, but from my research on it regular consumer TVs suffer from the same problem. If anything I'm more limited by the CM4's output, less the display itself.
Tradeoffs: I wanted to control my stack with minimal cabling pretty much everything else be damned, and this was a nice integrated solution and a excellent alternative to just buying a 43" dell monitor to hang on my wall. There are probably some image processing "features" that videophiles can comment on, but I've been used to literal decades of LCDs at this point, so an IPS display is going to be just fine(tm). Display lag is a bit higher than I would like for retro video games on it, but that's more an overall LCD problem, less a specific display issue.
In functionality it's much closer to a computer monitor than it is to a TV. The menus are obtuse, brutalist even, but if you ever had to work in IT and set up a conference room display you will feel right at home.
It's also a heavy pig compared to consumer televisions, right on the edge of what my mount is able to hold, and not really great to wrangle in a position by your lonesome, but not impossible either.
As for video: It's more than fine. Maybe the pixels are bigger than my macbook's screen, but at 6+ft viewing distance with my eyesight, I can't tell.
It would be nice if the article cited the actual contracts the labs have signed so that others could also read them and draw their own conclusions (maybe it does and I got tired of the slop-like writing style too early?).
Are the contracts actually take-or-pay-style? What are the terms? What are the amounts of compute and money involved for each future time period? What happens if the datacenter costs spiral upwards? What happens to the datacenter investment if the buyer goes bankrupt, goes public, or gets acquired (potentially by the datacenter owner)?
I personally think the datacenter build-outs are going to generally be a big swing and a miss. The problem 1-2 years ago was making models good enough to be useful for a variety of tasks. Now we have that. The next problem is making the models efficient enough to run a profitable business. Recent Chinese lab model releases (because they're already constrained on compute resources) and OpenAI price cuts on Luna seem to indicate that this transition to chasing efficiency has already started.
> The next problem is making the models efficient enough to run a profitable business.
In that world, what does Oracle do with a bunch of data centres which 1. It needs to pay for and 2. nobody needs. This is what the article is about: building supply far in advance of demand.
I don't know, but unless that contract is public, not many people know. I get the impression Oracle got a little bit of market pushback on the risk, as theirs seems to be one of the largest.
What contract? It's mostly corporate debt. Seems pretty plain and simple to me.
The contracts I guess you're talking about are compute obligations which frontier labs have. I don't need to know what those obligations are to know what happens to Oracle.
Contrast with Claude Government with all green 100% uptime. So either no one is using it or the people within Anthropic running that one have different priorities.
1996 was also the launch of APS-C which had the same ratio. Curious if Kodak just reused existing hardware for the camera so matched that with the sensor?
I think that's a different question; i.e. the ratios of 756x504 and 720x480 are both 1.5:1 (or, as we say in the camera world)
IMO the most likely factor for camera makers (and/or sensor suppliers[0]) is 'how much resolution can one pack into a sensor for marketing purposes while keeping yield margin in check for profit'. Megapixels instead of Megahertz...
[0] - Maybe different back then but the current state is that even the 'big names' will get their sensors from 'competing' camera makers...
In the USA step one would be to somehow build like 30 million houses so that the 10% extra rule could be established. That is an immense undertaking. Totally doable but it will take more than one government administration so has effectively zero chance of success.
That alone would probably be enough of a relief on housing that we wouldn't need to take any of the more drastic measures, and it effectively costs nothing fiscally to implement.
reply