I have to be very care buying spaghetti, because I have to carefully avoid tubular, thin, instant, spaghettoni, etc. As well as fettuccine and others because they're in similar packaging. My family is distraught if they get served anything other than the standard no.4 spaghetti!
I'm not the biggest fan of angel hair either, but one thing that it does have going for it is that it cooks about as fast as fresh pasta, like 2-3 mins. It's no bucatini but in a pinch or with small hungry humans to feed it's nice to have around.
Exactly this! And you have to expect that as the diameter approaches zero your pasta becomes indistinguishable from a gelatinous lump of gluten aka paste.
Our record at the office was 47 minutes. FORTY SEVEN MINUTES keeping a scammer on the phone till they hung up. We never laughed so much at work in all of twenty years. :D
I've had a similar experience for a local job. I honestly thought it was my previous employer unknowingly reaching out to me initially through a recruiting agency. The job description was verbatim my resume. Got the interview, and ghosted.
My guess would be using "AI" to increase/enhance sales with your existing processes. Pay for this product, get 20% increased sales, ad revenue, yada yada.
Indeed, you can strip out a whole host of things from the GPU, the framebuffer, the Z-buffer, the transform and lighting engine, instead filling it with more CUDA cores and a higher bandwidth memory controller with a larger bus, etc.
It's already there. Have you seen the six figure AI chips that Nvidia is selling to the data center customers? Those chips are no GPUs, they can't draw a single triangle or map a single texture, they're AI accelerators all the way. People still think Nvidia is selling gaming GPUs for AI workloads like it's 2018?
Google, Meta, et-all are working on their own AI chips but those chips will have to beat Nvidia's at Performance and TCO and Nvidia shows no signs of slowing down to let competitors catch up.
The chips are optimised for matmuls, but not for transformer architecture per se. With dedicated ASICS, and weights hardcoded (or stored in SRAM) we could theorically get 1 token per one cycle - so millions/billions of tokens per second, not hundreds.
Etched, for example claims they have a chip reaching 500k tok/s in the works. Which is still far from the theoretical max with the current techology.
A similar scenario went with Bitcoin's GPU/FPGA/ASIC - the current ASICs are millions of times faster than GPUs.
That’s fine if you never need to improve the model, which is valid in some use cases, but for chat style interaction or even code generation you’ll regularly have to update the weights.
Depends on a chip architecture - etched claims 0.5M tok/s with weights that can be updated. The main constraint is with the model architecture, where it needs to be specific transformer-based model. But they claim the chip can do both Mixtral and Llama - so the constraints are not too stiff.
TCO, yes. Raw performance, not necessarily. TCO will attack NVDA's margins. When Meta last wrote about their cluster it was presented as power equivalent to X NVDA chips. They are already bringing their own chips into the mix.
With Bitcoin I feel like it’s different, since the hashing algorithm would only ever change during a fork. This is rare in that it only ever happens every few years.
With AI, we’re constantly training different models, which can’t be trained using asics. If we ever get to the point where we no longer need to train new models, then yeah, it will go the way of bitcoin.
> With Bitcoin I feel like it’s different, since the hashing algorithm would only ever change during a fork. This is rare in that it only ever happens every few years.
Wait what!? Did the Bitcoin hashing algorithm ever change?
The problem with this comparison is Bitcoin has basically just been SHA256 for 15 years and likely will continue to be for some time.
Transformers have been mostly dominant for at least several years but there are still other archs (CNN, RNN, etc) in various use-cases and we're already seeing nearly-fundamental changes in Transformers and "emerging" approaches like Mamba, RWKV, hybrids, etc. Transformers have shown remarkable versatility and adaptability (that's their whole thing) but it's already creaking and showing its age.
Startups building Transformer-specific silicon are playing a very risky game that is already somewhat problematic now and almost certainly won't end well.
AI is much newer, much more vast, and moving much more quickly. The ASIC design, tape out, manufacture, software ecosystem, actually getting to market, etc cycle is fundamentally too long and I suspect even the Transformer-specific silicon we see now will be viewed as a major blunder in the relatively near future:
"Oh yeah, remember those graveyard companies that did transformer silicon back in the first AI hype round?"
I cannot see how anything other than GPGPU, TPU, NPU, etc (or similar "generic" approaches) will have legs.
This is one of the first questions I ask when buying a car. Do you put that stupid badge on everything you sell? I make it clear no sale if they do. I bought the car, you made your money, I'm not advertising for you, end of transaction.
Higher end German cars typically have a badge delete factory option. I have an older Porsche that came without badges from the factory.
I also think it is in poor taste to have prominent logos or brands on a product… if the manufacturer is going to use me for advertising that needs to be explicitly the deal: they provide the product free, and then we can negotiate how much they will pay on top of that.
Oh, some of those are the _worst_. My wife bought a car, and the dealer had gone beyond the usual decal and license plate holder, and had put an insert in the third brake light. E.g., every time she touches the brake, it's flashing this dealer's name. Clearly, this was not going to fly, but I was amazed that they tried.
Ahhh… yeah I would not tolerate that either, but I don’t buy cars from dealerships- they are middlemen in the business of ripping you off. They already had to buy it at a market rate and added no value, so can’t make anything unless they overcharge you.
I think that's an oversimplification. The market rates for wholesale, private sale, and retail are different. Selling a vehicle to a private party always nets the seller more cash than selling the same to a dealer. As a private buyer, you would have a very difficult time trying to buy a vehicle direct from most manufacturers.
As another said, this is due to lobbying, but that doesn't change the fact that the markets are different.
Fair point for new cars, I neglected to state I was only talking about buying used cars. Personally, I buy used cars because the amount of depreciation makes them an incredibly good deal, it seems to be an irrational cultural thing. I find it amusing that I get comments like "I didn't know you were so rich" for driving an old Porsche Boxster I got for $8k on Craigslist... always stored indoors and looks like new- and has been as reliable as a new car in 4 years of commuting. A new low end economy car at 3x the price is way too expensive for me.
Most individuals selling a used car, especially wealthier ones, aren't too interested in getting the most money from it, but just want it gone to move on with their lives. Not so with dealers, if they can't get you to pay a huge margin over what they bought it for (a price you could also buy one for) they'll wait for another buyer.
I'm not fundamentally opposed to buying used cars from a dealer, but when I've researched fair market prices, I've usually not been able to talk them down to anywhere close to them... whereas I've had many times where private sellers took an incredibly low ball offer I wasn't really expecting them to take.
When it comes time to settle up the car purchase just adjust the total for the monthly rental of advertising space on your car. Say, $200/month, 5 years in advance?
