woadwarrior01

7 ms·

91.
by woadwarrior01 2mo ago
> cc-by-nc is a protection against that. I don’t think so. There are CC licenses that allow for commercial usage. Most companies ban their developers from using anything with an NC or GPL style licence (with an exception for the Linux ke
92.
by woadwarrior01 2mo ago
Blink to approve. :)
93.
by woadwarrior01 2mo ago
I think the fully autonomous weapons systems restriction is not a very serious one and is very easy to circumvent. What if you can hire a human to push a single approve button? And what if that human's job is to front load approvals by
94.
by woadwarrior01 2mo ago
Why extend an MIT licensed model's weights (BiRefNet) and release it under a cc-by-nc-4.0 license?
95.
by woadwarrior01 2mo ago
Since the model is natively MXFP4, I think it'll be even more interesting on the hardware front. It'll comfortably fit on a 8x AMD MI355X node. I suspect that'll drive token prices down, further.
96.
by woadwarrior01 2mo ago
You might want to look at Peewee's query filtering syntax. https://docs.peewee-orm.com/en/latest/peewee/querying.html#f...
97.
by woadwarrior01 2mo ago
Great to see ruff, ty and uv being actively developed, even after Astral was acquired by OpenAI.
98.
by woadwarrior01 2mo ago
aka Baksheesh in the Middle East
99.
by woadwarrior01 3mo ago
I do this, but with Post-it notes. Something to do with the deification of books in the culture I was brought up in, which makes writing on the margins of books a sin.
100.
by woadwarrior01 3mo ago
Also, people seem to be oblivious to the existence of the neo-cloud: Nebius whose services and servers, a lot of western AI companies use. Nebius was spun out of Yandex 2 years ago to assuage the "supporting Russia" allegations an
101.
by woadwarrior01 3mo ago
Krea-2-Turbo and Z-Image-Turbo, both offline.
102.
by woadwarrior01 3mo ago
Thanks for bringing this up. You're 100% right. But most people, even technical ones are oblivious to how much of a difference modern samplers and higher quality quantization algorithms make for on-device LLM inference and are stuck wi
103.
by woadwarrior01 3mo ago
It is a reasonably non-opinionated model. My usual test is to ask these models to generate comic book and cartoon characters that hosted image generation models refuse to generate. I think that text is just CYA legalese.
104.
by woadwarrior01 3mo ago
I've seen this pattern repeated in a number of large production codebases using optional booleans.
105.
by woadwarrior01 3mo ago
Krea-2-Turbo. I've even got it working locally on my M5 iPad Pro.
106.
by woadwarrior01 3mo ago
Ironic that the app is named Nativ(e) and yet bundles a full Python runtime. Nonetheless, still less bloated than LM Studio, which bundles a full Python runtime and electron.js (which in turn bundles a whole browser runtime).
107.
by woadwarrior01 3mo ago
Works with the Bonsai-27B-1bit-mlx model on my mac. Edit: this is running with MLX, and not with WebGPU on the browser, also it is a fail. """ You should *walk*. Here’s why: - *Distance*: 50 meters is very short — about 0.3 m
108.
by woadwarrior01 3mo ago
I think the "8.9 million developers use ollama" figure they throw around is contrived. I found this 2025 article which posits that there were 47.2 million software developers in the world. Even the number of developers has doubled
109.
by woadwarrior01 3mo ago
To me, it's the cli.
110.
by woadwarrior01 3mo ago
It's unlikely that a clear single winner will emerge in this competitive market.
111.
by woadwarrior01 3mo ago
Yeah, I learnt about this from the amazing Max Bennett book: A brief history of intelligence.
112.
by woadwarrior01 3mo ago
BaaS: Bhatta as a Service. nb: Bhatta is Sanskrit for priest.
113.
by woadwarrior01 3mo ago
I think NERVA and its Soviet equivalent RD-0410 were much more practical and plausible. Unlike Orion which was complete vaporware, both these projects reached the ground test prototype stage. https://en.wikipedia.org/wiki&#x
114.
by woadwarrior01 3mo ago
Next update? Have you read their privacy policy? https://lmstudio.ai/app-privacy
115.
by woadwarrior01 3mo ago
> In any case I am not sure pivoting from running local models to "cloud offering" (as in providing llm inference at their severs) is a sensible choice granted there is already competition in that space and they have no leverag
116.
by woadwarrior01 3mo ago
The desktop app's GUI is electron.js based.
117.
by woadwarrior01 3mo ago
Ultimately, the onus at every VC backed local LLM startup is to launch a cloud based offering, because that's the only potential path in sight for venture scale returns.
118.
by woadwarrior01 3mo ago
Small encoder-only transformers are excellent at classifying LLM-Generated Text. I built an on-device iOS app using a custom small encoder that achieves an AUROC of 99.81 on RAID-bench.
119.
by woadwarrior01 3mo ago
Cursor will rebrand it as Composer 3.0 to assuage any such concerns, as they did with the previous Kimi models.
120.
by woadwarrior01 3mo ago
$56 premium for the OpenAI skin. :)

More ›