alfiedotwtf

7 ms·

91.
by alfiedotwtf 4mo ago
Unfortunately, yeah. > The main issues were related to write speed but we do a compaction of the database and write performances are way better after that I'll ping you if I ever get around to rewriting a faster kv-store in Rust :)
92.
by alfiedotwtf 4mo ago
The management to handle Threading and IO and not get it wrong is pretty high unless you’re using something like Rust because it guarantees you’re safe, vs letting the routine runtime management handle everything for you for free
93.
by alfiedotwtf 4mo ago
> And if compiler authors disagree, they are morons I remember arguing with Howard years ago on “C vs Rust”. He said that you don’t need Rust, you just have to be good at C programming, so I pointed out CVEs in LMDB attributed to his own
94.
by alfiedotwtf 4mo ago
How’s LineageOS compatibility these days? And besides F-droid, is there a place where mobile apps are plentiful without being full of malware? Also, how’s isolation on LineageOS for mobile apps? I think I’m getting to the point where I’m th
95.
by alfiedotwtf 4mo ago
Yes! But as a Plan B, why aren’t we emulating Android on these devices (or is it the Secure Enclave that’s the spicy bit that these apps need)?
96.
by alfiedotwtf 4mo ago
> Each of these companies relies on a deep network of suppliers Isn’t this by design? ie each country customer has an important stage in the supply chain so that if the primary country fails to deliver, a supply countries can choke globa
97.
by alfiedotwtf 4mo ago
If you squint hard enough, it sounds like their storage layer is a bloom filter
98.
by alfiedotwtf 4mo ago
... and under an hour before the first phone camera recording hits twitter
99.
by alfiedotwtf 4mo ago
You haven’t lived life until you’ve sat a Data Structures and Algorithms exam having to write all coding solutions with paper and pen
100.
by alfiedotwtf 4mo ago
> Not trivial, you would need to do lots of evals and prompt tuning when you switch models. Couldn’t we just train smaller models to “translate” what the harness user wants to what the worker model expects? I mean, if models understand c
101.
by alfiedotwtf 4mo ago
> CC silently alters the system prompt using invisible-ish Unicode markers. It encodes proxy / gateway classification into a sentence that looks like plain English. It hides the domain list behind XOR and base64. This is not a malic
102.
by alfiedotwtf 4mo ago
GLM 5.2 is the elephant in the room. GLM 6 will probably be a Claude Killer
103.
by alfiedotwtf 4mo ago
For how long lol
104.
by alfiedotwtf 4mo ago
So that’s for building muscle, but what about if you wanted to lose a few kilos and increase endurance for long distance running? What would be the way to go to optimise your time?
105.
by alfiedotwtf 4mo ago
And before that, graphics programmers called it vertical retrace :upsidedown:
106.
by alfiedotwtf 4mo ago
> Similar to compiler engineers before. I guess the difference here being that we have ample compiler literature and practically know 99% of all there is to know about compilers that exist in the wild vs this new field. Until we’ve gath
107.
by alfiedotwtf 4mo ago
Qwen 3.7 120B will kill off Antropic’s IPO
108.
by alfiedotwtf 4mo ago
Step 2.7’s MTP seems broken (at least for ik_llama.cpp) where the draft model starts and ends in block 3 but ik_llama bails out looking for block 0 :(
109.
by alfiedotwtf 4mo ago
Fords have a really bad name in Australia. The cheapest Hyundai will outlast even a high end Ford. But in terms of reliability, here Toyota is king
110.
by alfiedotwtf 4mo ago
@ihsw you’re currently greyed out so I can’t reply, but lol holy shit i’d be updating my resume if i were you
111.
by alfiedotwtf 4mo ago
$19 Chinese IP cameras? Try $800 dashcams. Blackview dashcams by default join the company’s public webcam feed which is viewable inside their app.
112.
by alfiedotwtf 4mo ago
Oh! Looks like I’ve been sleeping on Laguna!
113.
by alfiedotwtf 4mo ago
If you’re able to run qwen3-coder, have you thought about 3.6 27B or 35B? Looking at benchmarks, 3.6 looks its gained a lot over qwen3-coder
114.
Ask HN: Running local LLMs? What's your model and hardware
11 points by alfiedotwtf 4mo ago|9 comments
115.
by alfiedotwtf 4mo ago
> By sitting in the alcove, and keeping well back, Winston was able to remain outside the range of the telescreen, so far as sight went
116.
by alfiedotwtf 4mo ago
> source more and higher quality (mostly synthetic data) Kind of an oxymoron don’t you think. If they could generate data that looked kind of real, why don’t they just generate that data on the fly during inference
117.
by alfiedotwtf 4mo ago
> And the chinese labs also have incentives to keep releasing models Not really.
118.
by alfiedotwtf 4mo ago
Exactly my worry. I’m optimistic in the future the EU, the EFF, the GNU, or the Linux Foundation could have been the umbrella to run a LARGE open model for everyone. It’s sad to think that Mozilla spent years and millions doing virtual real
119.
by alfiedotwtf 4mo ago
Yeah, I was talking about recovery. I’ve always wondered if systems could be compressed via trying to curve fit generators like these oscillators. … I still hold hope metaheuristics could at least do a good job of approximating. One day I’l
120.
by alfiedotwtf 4mo ago
$100 Million to The Trump Foundation, and Anthropic get to become the US AI Regulator

More ›