_davide_

in reality even the mention of a prohibition is enough to make the model reject that no matter what

i beg to differ, in an ideal world a system possibly is a binding law and high end models are starting to be really aligned to the exact system prompt. The instructions must be simple to follow, if you start doing complex rules it'll call apart, but I'll usually follow the stringer interpretation.

or completely deleting the app and use the browser version for desperation

Disagree with the conclusion, even without carefully curated context every high end LLM perform just as well, maybe with an extra detour. In contrast if even one of the learnings is not up to date or doesn't apply to the current situation you find yourself with a long detour or even a failure.

Yes, given a direction, they are pretty good at "fuzzing", trying out everything and eventually find something. Super useful, but it doesn't have the same level of precise targeting you could expect from a high end model.

Just go online rent a few servers, download K3, ablate it, run the thing, shut it down. Will probably cost a few hundreds bucks to ablate, but f* this non-sense paternalistic sh*.

I wish i had the time to do it and blog it, "in your face" kinda style.

this is a silly oversimplification

there is a simpler way, make a complete mental model of the changes and ask questions to confirm your understanding. so much faster.

feeding an existing conversion with output and reasoning from model x to model y will have many side effects, most likely a degradation of perfomance and alignment. should be done at least within the same family and generation of models... I'm not a huge fan of routers

php is still a thing?

What's the point of "releasing it"? It only makes sense in the labs, and what would calling your APIs give "me" as a researcher, other than a baseline to beat? mah the model is real, but it feels 100% internal

Seriously ready to fully drop google emails and services, i'm using vivaldi + personal email for a while now, the migration was slow a bit painful but complete. Only the phone is still attached to play store and google, but i'm 100% ready to switch to the fully supported Chinese variant of ColorOS and flush Google into the toilet once and for all. But my next phone will 100% be a Sailfish again.

pls..? more like `if you don't we'll all die, so do it`

This post seems to be created out of thin air rather than real experience and data

Loved this incremental evolution, things gets way more understandable...usually xD

After this level of bullshit I won't ever spend a second any other new/blog from anthropic

> Apple Silicon is a natural fit for tensor/matrix operations due to unified memory,

This is a common misconception. This is NOT true: memory designed for CPU performs horribly for GPU tasks. Mac/Strix Halo/NVIDIA Spark are performing horribly compared to a desktop video card with GDDR.

I spent quite a few weekends optimizing ROCM kernels for strix halo, i WISH I had GDDR instead of high-frequency CPU RAM; the bandwidth would be SO MUCH better.

I'm quantizing weights not to make computation faster, not because I'm out of memory, but because memory cannot move fast enough to be processed and it's cheaper to load quantized version, convert into bf16 compute, and discard; This happens every single time a token is generated for the whole model over and over.

him being nazi was inconsequential, the richest person in earth being Nazi is.

> Try separating politics from the advancement of humanity, you'll feel better.

can't. won't.

just subscribed

I'm do so as well, i tried qwen3 omni 3 but it was ridiculously stupid, and i ended up with stt thinker and tts. kokoro for now

lol, keep hoping

i meant that even if you are okay with giving up privacy, the bare minimum accountability is missing from police, so it isn't really an option

To balance it, the police need to be extremely accountable, but so far they get away with murder pretty easily...so...

I'm using my own agent, but i can't risk blocking the company account with it.....

for lack of directonality?

If compute is not the bottleneck, memory is easy-ish to produce (the hard part is mostly on the fab side); what stops a Chinese NVIDIA (huawei) from being 10x cheaper?

They are usually the same family, LPDDR is used for amd and macs, but the fabs are the same as the most expesive HBM memory, if they have a choice they are going to produce the ones that they can sell for more $$.

I'm writing my own inference engine for Strix Halo and the same model. I already have 30%+ performance plus a more graceful decay over long contexts; that said, their point stands: memory bandwidth is what you really want.

same experience here, as soon as it touched any gpu code it stopped working

> This is very literally what already happens, it's called a EULA. Yes, but they "reserve the right" to update whenever, making it pointless

> "In favor of the customer over anything else" is not a legally viable clause. I'm sure that legislators could put the principle down in a much clearer way. What's lacking is the will.