Infernal

Did it used to be a lot bigger? Looks like a couple wooded lots side by side now.

I am really surprised you’re seeing half the generation performance of 3.6 with 3.8 at the same parameter size and quantization (and same prefill performance to boot) - is there just an optimization in the stack somewhere for 3.6 that hasn’t landed yet for 3.8?

Curious if other folks are also surprised by this apparent discrepancy or have a ready explanation.

If I understand correctly, it’s not that the LLM can’t write a good comment, it’s that you want to be able to interpret and understand the generated code without comments - and in that process end up writing comments yourself.

Core Series 3 processors, Thunderbolt 4, a new keyboard with fingerprint reader and backlight.

As a current Framework 12 user I'm not in a huge rush to upgrade the mainboard, but will be snagging the backlit keyboard as soon as it's available by itself. Would love to see a haptic trackpad one day, as well as a screen that covers a larger color space.

Ooo which 80's sedan? I had an E28 20 years ago that I still think about

Not just you - from ansel.photos:

>Ansel has not published a stable release yet and there is no ETA for one : a new release is published when the list of all bugs have been cleared, so we know the software and stable. So far, Ansel only publishes revisions, which are intermediate states of the sourcecode.

I have a similar question and I’m inferring the answer is no - look at the cache hit rate of 23% for the 128GB M5 Max. I had previously assumed that the 40B active meant that a set of layers was chosen as THE expert for a given prompt and generation was then limited to those layers until complete. But in that case you’d have expected the expert caching to have a super high hit rate once you had enough RAM to hold an entire expert’s worth of layers.

Do we know where those key players went?

If I’m not mistaken it appears the regular S100X has these too. The only difference really seems to be the lacquer finish, and maybe the navy display.

On top of that, they’re all sold out and the cheapest I can find any reseller is $1200+ USD

I didn't even know there were 4 methods - supposedly Swiss Water Process is the best in terms of not affecting the flavor or exposing you to exotic solvents, is one of the four superior to SWP?

So I thought your initial comment was a (pretty good) joke about using a microwave oven, but now I’m not sure. Is this testing license you reference a continuation of the joke or a real thing?

I’m searching for the source material here but must not be using the right keywords, any hints?

40's as well but first line is basically, hey here's some MIT licensed code we can do whatever we want with it, surely we can use it to make some money to the point we can afford private jets in and out of Dubai.

"TeamViktor's faces surely?" = imagine the looks on their faces "diamond handsing to the moon" = hanging onto a situation that looks bad now hoping it gets better, throwing good money after bad, the opposite of cutting your losses.

Is there an article about this written by a human? The “it’s not X. It’s Y” is too distracting.

Your car wouldn’t happen to be a Volvo would it?

For anyone stopping by looking for more info, it’s Camp Lejeune not Legume.

I'm not sure where else you can get a half TB of 800GB/s memory for < $10k. (Though that's the M3 Ultra, don't know about the M5). Is there something competitive in the nvidia ecosystem?

On some vehicles it’s easier than others. Unfortunately it’s a great idea to research before making a purchase decision.

Oh man, my grandmother was like this with finding four leaf clovers. She would just find them constantly, all the time, on command, or maybe while standing around having a conversation. Her description of it was "it's like they're just jumping up and waving at me" which somewhat fits with the author's description of motion. Never heard of anyone else like this though, neat to see others in the comments.

Fair, I suppose I've misunderstood. I took "it's been available since 2024" as a dismissal of this new information.

What does the GPD Win 4 do in this scenario? Is there a step w/ Agent Organizer that decides if a task can go to a smaller model on the Win 4 vs a larger model on your Mac?

The point here is that the doc you linked is a year and a half old, this (if real) is much newer. Security is a constant arms race between attackers and defenders, nothing is static so updates of this nature are always welcome.

> one that's a device I actually own and control.

Where do you find such a device today? Pixel with GrapheneOS? Any decent mobile Linux hardware I don’t know about?

But that passphrase you saved is an additional key, in case you lose all your Apple devices for example. You can tell it isn’t required for your phone to decrypt data because you don’t have to type it in to access your data, or even migrate to a new phone.

> Or will the iPhone have a multi-hour update where it decrypts its entire iCloud archive on the client-side, and then reuploads it without encryption?

More likely that the phone just sends the keys to Apple in that case

If I understand correctly it doesn’t matter where the user’s profile is hosted, the point is the user has a store of data that can be accessed by multiple apps via the AT protocol, only naming convention separates one app’s data from another’s within a given user’s profile.

TFA says “The facade’s cement chevrons catch the sun” but I’m not sure about the ones inside.

AI

I think the comment you're replying to agrees with you - it makes more sense with the last two quoted sentences swapped. I could be wrong but this interpretation seems consistent with the rest of the comment.

"Many software companies hire smart people as CPO, Product Director, or Head of Product because they have leadership skills, people skills, and some knowledge of the industry. Some companies go even further and promote an engineer with project management experience to Head of Product. But most have little to no background in business, marketing, economics, or product development."