_davide_

4 ms·

31.
by _davide_ 2mo ago
pls..? more like `if you don't we'll all die, so do it`
32.
by _davide_ 2mo ago
This post seems to be created out of thin air rather than real experience and data
33.
by _davide_ 2mo ago
Loved this incremental evolution, things gets way more understandable...usually xD
34.
by _davide_ 2mo ago
After this level of bullshit I won't ever spend a second any other new/blog from anthropic
35.
by _davide_ 2mo ago
> 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 ar
36.
by _davide_ 2mo ago
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.
37.
by _davide_ 2mo ago
just subscribed
38.
by _davide_ 2mo ago
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
39.
by _davide_ 2mo ago
lol, keep hoping
40.
by _davide_ 3mo ago
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
41.
by _davide_ 3mo ago
To balance it, the police need to be extremely accountable, but so far they get away with murder pretty easily...so...
42.
by _davide_ 3mo ago
I'm using my own agent, but i can't risk blocking the company account with it.....
43.
by _davide_ 3mo ago
for lack of directonality?
44.
by _davide_ 3mo ago
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?
45.
by _davide_ 3mo ago
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 $$.
46.
by _davide_ 3mo ago
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.
47.
by _davide_ 3mo ago
same experience here, as soon as it touched any gpu code it stopped working
48.
by _davide_ 3mo ago
> 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 vi
49.
by _davide_ 3mo ago
Yep, that's me. the only real blocker is that American companies don't trust Chinese providers, but i could just find a good American provider that hosts DeepSeek and/or GLM. I would at least be able to choose my own agent i
50.
by _davide_ 3mo ago
A simple law: everything the customer buys must always behave *in favor of the customer over anything else*. If the product/service contradicts this, it must be fully stated before the purchase and cannot be updated. <= This would b
51.
by _davide_ 3mo ago
I'm tempted as well, just out of spite
52.
by _davide_ 3mo ago
It isn't a promotion, it's 2x the parameters of opus and we are paying with 2x the consumption rate. They just want to get rid of the subscription model.
53.
by _davide_ 3mo ago
"promotion" like they are doing you a favor just this once out of their goodwill... Really really really pissed me off
54.
by _davide_ 3mo ago
What a well written article!
55.
by _davide_ 4mo ago
just buy more RAM, it's cheap enough...
56.
by _davide_ 4mo ago
the threat is non-existing for agentic flows. Local interfere could catch up on high end phones
57.
by _davide_ 4mo ago
Sounds like a good solution my Führer
58.
by _davide_ 4mo ago
you can absolutely use it for some workloads, but as soon as you have some extra complexity for a big repo it'll take forever and the economics are so silly to the point that the electricity bill would be comparable to a subscription.
59.
by _davide_ 4mo ago
i used to mix remote and local minimax 2.7(q3) on my strix halo, it run at 30 tg and 220 tokens pp... it was a bit painful slow, but it was a good feeling i could stay offline. unfortunately m3 which is at opus .8 levels is 460b parameters
60.
by _davide_ 4mo ago
I did develop my own agent around MiniMax. I did see weird behavior when I messed up the loop, like omitting pieces of remove thinking; maybe it's an agent bug, some models/providers just ignore/normalize the broken input, so

More ›