

The spectrometer knows where it is, because it knows where it isn’t.


The spectrometer knows where it is, because it knows where it isn’t.


Unlike conventional spectrometers, which rely on dispersing light or algorithmic reconstruction to recover spectra, the convolutional spectrometer physically performs a convolution operation on the incoming light. This is achieved using a simple cascade of optical components with periodic spectral responses, such as unbalanced Mach–Zehnder interferometers or micro-ring resonators.
Is this thing a spectrometer or a fucking turbo encabulator?


The more you tighten your grip, Google, the more ad revenue will slip through your fingers.


Rail lines use multiple redundant navigation and safety systems. GPS is one of them.


It frightens the ignorant white middle class that The Heritage Foundation has been grooming for decades to be compliant through fear.


“We can’t manipulate this data to support our preconceived notions. Clearly the problem is that we don’t have enough data.”


Yeah, I think we’re talking about the same thing. I thought the article I linked was the one I had read about model inbreeding, but now that I look at it a bit closer, it’s probably the product of model inbreeding itself. ;) I thought there was an article published this year about the problem, but now I can’t find it to save myself. It’s possible that I’m hallucinating. My memory is worse than ChatGPT’s.


I’ve been a software QA engineer since 1991. I wound up working as a professional mover during the dot com bust.


Because there will be so many agents, the world will not be limited by the number of people
> The world will need more developers
> The world will not be limited by having fewer developers.
Choose one.


“Biomass tariff. Just in case you take it back to America with you. If you poop in Italy, save your receipt, you’ll get a refund.”


AI is the new dot com. Pity it’s going to take so much of the global economy with it when it sinks, but the sooner it happens, the less damage it will do.


I’m glad to hear you’ve found such exceptional stability, but the norm is closer to 1-2 years.


AI developers might be smart enough to filter their own output, but AI training is still taking in output from their competitors’ models, usually without realizing it.


Software has never been considered stable, as the article mentions in the first paragraph.


AI is already training on slop, because it can’t tell its own excrement from actual human creativity.


This. It’s not a plan, it’s an excuse.
Yeah no it won’t. The only uncertainty is whether it’s worth alienating customers to save a few bucks on performers, now that you can’t get away with it any more.