I’d agree there’s no single right answer, but I think generally a risk profile for an insurance product that is so outsized against you seems like a poor decision. This isn’t like putting all your money in meme stocks…
Total loss being the bar seems crazy to me. A 20-50k unexpected expense for almost everyone would essentially result in bankruptcy or at minimum insolvency and long term debt. That means you would need to hold enough…
What do you consider catastrophic loss? I’ve found in home insurance the delta on lower deductibles isn’t that meaningful and when something happens, the pay out is almost always worth it. Something like water damage…
Again I see this as unrelated. You can be passionate about the job, but the location is a checkbox.
I see it a little different. I'm not saying I prioritize work over everything else, I just see them as almost unrelated. The location is a filter, not a motivation. For various reasons, I wouldn't ever choose certain…
> Most people want to have a job set up before they move to a new city. This is obvious for most people... generally a bad financial idea to move to a new city without a job. But, primarily taking a job just to move to…
Maybe I just haven't met these engineers, but isn't the problem usually "I don't want to relocate" ? Who talks about relocating as if that's the primary motivation for a job?
> Annoyingly the input pins had a constant DC offset of over 2 volts, which went away when shorted to ground but reappeared moments after I removed the short. Mic inputs on audio codecs always have DC bias, because they…
So, even worse than 0, if we’re saying the value of games is not measurable but substantial?
> Or they could create 1000 users in a row with 4GB passwords and fill up your storage. Not an issue if you’re storing passwords properly. You don’t store the actual password, you store a (hopefully) salted hash of it.
You’re not replacing the outsourcing it component though, you’re replacing a maid at home doing it for you. In home laundry services are a very different experience since you don’t have to also go pick up and drop off…
Is there a reason head gestures couldn't be mapped to functions in Linux? I see little reason why you couldn't have it used to play/pause or something. Heart rate monitoring less-so... but even that exposure could be…
I think you’ve got things quite backwards if you think that the desire to run models on device or use any of the variety of open weight models (big or small) on premise is somehow bowing down to tech billionaires. Quite…
Why are you representing this as such a binary here? For SLM we don’t need the Taalas stuff at all. Just run it locally on your own device if it’s truly a small model. And there’s plenty of larger models that can be run…
> They expect to scale up to real LLMs by splitting a model layer-wise across several chips, which they can do without incurring any throughput penalty. I’ll patiently wait to see this in reality. Their demonstration…
But I’m not missing the point. If you can run one frontier model at 750t/s, then you can probably run many many instances of an SLM in parallel at a rate that exceeds 15k/s. That’s kinda the point of the flash or…
But it’s irrelevant. 750 tokens/s on a full frontier model is useful. 15000 poor quality tokens is much less useful no matter how much scaffolding you put around it.
It’s not just old, it’s also tiny and quantized. It’s llama 3.1 8b at 3/6-bit quant. This is the type of thing you can run on almost any device…
Or you could buy a MacBook Neo on various retailers who still have it for $599. Heck, it’s on sale for $589 on Amazon right now. I don’t think the market shopping for $600 laptops is the same market that wants something…
This is no different than them skipping the “Ultra” chips on some generations. The only real difference is it going all the way down to skipping the “Pro” line. So, only the MacBook Air, low end MBP, and maybe the iPad…
> I would not be at all comfortable, with a human server, making them wait while changing my selection multiple times (no I want it with the ginger sauce... no, without... no actually the sesame sauce... no actually I…
> Seems like a made-up distinction that shouldn't be necessary since M6 has not even released. Why would it? Each generation of the M series has an architectural improvement on their chipsets. The difference between an…
> All silicon shares the same process. CPU, GPU, SSD, etc No, no they don't. CPUs and GPUs do, broadly, but SSDs and RAM use specific processes catered only to their respective usages. You can't use a NAND flash process…
Until they raise the price on that too. Dell has explicitly stated it's a "limited time" price, so don't be shocked if it becomes the $699 XPS13 almost immediately.
> You really believe food, gas and house prices are not increasing at the same amount? No, I don't think food, gas, and house prices, have 5x'd in price like RAM has. This is abundantly obvious.
I’d agree there’s no single right answer, but I think generally a risk profile for an insurance product that is so outsized against you seems like a poor decision. This isn’t like putting all your money in meme stocks…
Total loss being the bar seems crazy to me. A 20-50k unexpected expense for almost everyone would essentially result in bankruptcy or at minimum insolvency and long term debt. That means you would need to hold enough…
What do you consider catastrophic loss? I’ve found in home insurance the delta on lower deductibles isn’t that meaningful and when something happens, the pay out is almost always worth it. Something like water damage…
Again I see this as unrelated. You can be passionate about the job, but the location is a checkbox.
I see it a little different. I'm not saying I prioritize work over everything else, I just see them as almost unrelated. The location is a filter, not a motivation. For various reasons, I wouldn't ever choose certain…
> Most people want to have a job set up before they move to a new city. This is obvious for most people... generally a bad financial idea to move to a new city without a job. But, primarily taking a job just to move to…
Maybe I just haven't met these engineers, but isn't the problem usually "I don't want to relocate" ? Who talks about relocating as if that's the primary motivation for a job?
> Annoyingly the input pins had a constant DC offset of over 2 volts, which went away when shorted to ground but reappeared moments after I removed the short. Mic inputs on audio codecs always have DC bias, because they…
So, even worse than 0, if we’re saying the value of games is not measurable but substantial?
> Or they could create 1000 users in a row with 4GB passwords and fill up your storage. Not an issue if you’re storing passwords properly. You don’t store the actual password, you store a (hopefully) salted hash of it.
You’re not replacing the outsourcing it component though, you’re replacing a maid at home doing it for you. In home laundry services are a very different experience since you don’t have to also go pick up and drop off…
Is there a reason head gestures couldn't be mapped to functions in Linux? I see little reason why you couldn't have it used to play/pause or something. Heart rate monitoring less-so... but even that exposure could be…
I think you’ve got things quite backwards if you think that the desire to run models on device or use any of the variety of open weight models (big or small) on premise is somehow bowing down to tech billionaires. Quite…
Why are you representing this as such a binary here? For SLM we don’t need the Taalas stuff at all. Just run it locally on your own device if it’s truly a small model. And there’s plenty of larger models that can be run…
> They expect to scale up to real LLMs by splitting a model layer-wise across several chips, which they can do without incurring any throughput penalty. I’ll patiently wait to see this in reality. Their demonstration…
But I’m not missing the point. If you can run one frontier model at 750t/s, then you can probably run many many instances of an SLM in parallel at a rate that exceeds 15k/s. That’s kinda the point of the flash or…
But it’s irrelevant. 750 tokens/s on a full frontier model is useful. 15000 poor quality tokens is much less useful no matter how much scaffolding you put around it.
It’s not just old, it’s also tiny and quantized. It’s llama 3.1 8b at 3/6-bit quant. This is the type of thing you can run on almost any device…
Or you could buy a MacBook Neo on various retailers who still have it for $599. Heck, it’s on sale for $589 on Amazon right now. I don’t think the market shopping for $600 laptops is the same market that wants something…
This is no different than them skipping the “Ultra” chips on some generations. The only real difference is it going all the way down to skipping the “Pro” line. So, only the MacBook Air, low end MBP, and maybe the iPad…
> I would not be at all comfortable, with a human server, making them wait while changing my selection multiple times (no I want it with the ginger sauce... no, without... no actually the sesame sauce... no actually I…
> Seems like a made-up distinction that shouldn't be necessary since M6 has not even released. Why would it? Each generation of the M series has an architectural improvement on their chipsets. The difference between an…
> All silicon shares the same process. CPU, GPU, SSD, etc No, no they don't. CPUs and GPUs do, broadly, but SSDs and RAM use specific processes catered only to their respective usages. You can't use a NAND flash process…
Until they raise the price on that too. Dell has explicitly stated it's a "limited time" price, so don't be shocked if it becomes the $699 XPS13 almost immediately.
> You really believe food, gas and house prices are not increasing at the same amount? No, I don't think food, gas, and house prices, have 5x'd in price like RAM has. This is abundantly obvious.