flockonus

4 ms·

1.
by flockonus 6d ago
Gotchu - https://jevmaxx.ing/
2.
by flockonus 6d ago
As a developer I've only updated my Mac around release date once. Faced so many bugs and incompatibilities that now I wait at least 2 months so both Apple & third party iron out the kinks.
3.
by flockonus 9d ago
No question OSS is amazing, but this video is a satire at best. It doesn't take much attention to see the results on right vs. left side are significantly different. Jev is not interesting if it's not "smart", a 1B param
4.
by flockonus 10d ago
Correct. Not to say we're getting into the weeds of probability here as well. "What are the odds a thunder will strike in Paris at 1pm UTC of 2026-09-16" - that could be a 0.001 chance going from blind historical measurements
5.
by flockonus 21d ago
"cyberattack" is indeed exaggerated. AI breakout risk most definitely isn't, specially given how their swarm did in fact hack HuggingFace not long ago.
6.
by flockonus 23d ago
It is exhausting to keep up with model releases yes, much like it was for a while during the Cambrian explosion of FE frameworks, eventually tech seems to work out to consolidation. But more so it seems there is Fear of missing out (FOMO) i
7.
by flockonus 24d ago
No idea what you're trying to say... he's been directionally wrong for years as the article illustrates clearly.
8.
by flockonus 1mo ago
Not really, not leading at any criteria, altho top 5 in most - https://arena.ai/leaderboard
9.
by flockonus 1mo ago
Agreed. During comp sci we got to re-implement various algos of networks, OS, database, firmware.. and it all gave complimentary intuitions that were useful when tackling practical implementations and bottlenecks.
10.
by flockonus 1mo ago
I knew the universe was nothing but a brick wall, now we have proof! https://viewer.legacysurvey.org/?ra=80.3728&dec=-69.4497&lay... /jk
11.
by flockonus 1mo ago
Depends who you work with, what's the intention, budget, etc. I'd agree this is a really good one. I'm used to incremental Figma wireframe -> final product and working together with a designer.
12.
by flockonus 1mo ago
There is some irony being a developer and reading along the lines of: "oh look at the comparison between these models executing a task for a few cents on a job i'd be charging 1k minimum"
13.
by flockonus 2mo ago
To the models you say for coding, can you give an example of what you code? I find them insufficient for my projects (mid sized), but curious what people see working.
14.
by flockonus 2mo ago
Oxigen is ALSO, and perhaps in vast quantities produced at the bottom of the ocean, by lumps of metal that react with sea water (aka Dark Oxygen). Interestingly, there are now massive marine harvesting operations targeting that very ore. h
15.
by flockonus 2mo ago
To establish oneself in a country (for the majority if not all the western world) without an appropriate visa is already breaking the law. Not saying this is the only way, or what should be - but it's what it is, selectively breaking t
16.
by flockonus 2mo ago
I do find the author position to be incoherent in the following regard: 1. believe a statement can be true or false, and false should be called out. 2. believe deportation is morally wrong in certain cases, that is, the law can be selective
17.
by flockonus 2mo ago
As you already went through the thought exercise of laying all this RAM over various slots, then match against the right CPU (which also you'll need multiple) - it becomes clear quite fast that it's trying to mimic the architectur
18.
by flockonus 2mo ago
Same here, as we're sitting in the middle between requests and what budget constraints are allowed given a particular token allowance there can be 10 ~ 100 milliseconds improvement in the UX (TTFT) given such massive tokenization speed
19.
by flockonus 2mo ago
Ah yes, i missed it because i cmd+F for "pric" & "cost" - but now i read more carefully, it makes sense. It also speaks about the biz model. Personally i wouldn't change a thing, other than perhaps add one of th
20.
by flockonus 2mo ago
At first i found odd that pricing is nowhere in the website, but then i found it burried under FAQ (free!) - is it to weed out people who are focusing on free rather than going deeper into the experience? > How much does attending the Re
21.
by flockonus 2mo ago
There is also the 1bit version @ 3.9 GB that retains 90% of the intelligence - quite a feat! https://huggingface.co/prism-ml/Bonsai-27B-mlx-1bit
22.
by flockonus 2mo ago
Excited to read if you ever get to publish! In particular 1. made me even more curious. I do get most of my AI value on prototyping [small codebase] & summarizing.
23.
by flockonus 3mo ago
Cool! I also remember from decades ago a friend in high school taught me "SOS" and still stuck with me: 3 short, 3 long 3 short, i suppose ... - - - ...
24.
by flockonus 3mo ago
> That was a departure from how I normally work since I did not write a single line of code. I will likely write something about my A.I framework (and opensource it) next month. If the author is around, super curious how they got to enjo
25.
by flockonus 3mo ago
> DOGE found an actually highly efficient Federal government Wish we could see the evidence for that.
26.
by flockonus 3mo ago
It would be true if there was a unified "Anthropic" entity making every decision from pure rationality. Instead, more tokens increase Claude Code team's metrics of token usage, which most likely has a KPI around token usage a
27.
by flockonus 3mo ago
I suppose it's like gradient descent, often times you get stuck in local minima and there is no inertial force to overcome the next peak.
28.
by flockonus 3mo ago
Curious for what an MTP only result would look like, both in terms of output quality & tk/s ?!
29.
by flockonus 3mo ago
Looking at this gave me a renewed (even if brief) sense of appreciation for our society. Even with all of its problems, for this moment in time our society is operating effectively enough that humans can engineer interesting, beautiful, cul
30.
by flockonus 3mo ago
Some irony in so many posts about AI becoming more capable at programming, at the same time, top post on hackernews is a game about where you code by reading a magazine like it's 1997.

More ›