vFunct

11 ms·

91.
by vFunct 1y ago
Studio lighting isn’t natural lighting nor is it bounce flash lighting. I am specifically talking about amateurish event/wedding photographers that use bounce flash. Which have a iPhone use commonality. (You never use iPhones in studio
92.
by vFunct 1y ago
Yah you can't fix color balance from bad color cast. ALL photos look good with direct flash. Never use bounce flash. And indirect lighting is never flattering. EVER. Fire any photographer that ever uses bounce flash. Nobody wants their
93.
by vFunct 1y ago
They're still over saturated. Skin tones always have a cosmetic/tanned look compared to real life. Mirrorless camera photos have a lot better output. You can see that even in the first sample comparison. If you look at the photo o
94.
by vFunct 1y ago
please please apple give us a M5 MacBook Pro laptop with 2TB of unified memory please please
95.
by vFunct 1y ago
> Most Israelis do not hate Palestinians. LOL 51% of Israelis gave Palestinians a 0 out of 100 score rating of their humanity, according to PCPSR and the Times of Israel poll: https://pcpsr.org/en/node/989
96.
by vFunct 1y ago
Seems like their business plan is unsustainable. What's a sustainable cost model? Say an 8xB200 server costs $500,000, with 3 years depreciation, so $166k/year costs for a server. Say 10 people share that server full time per year
97.
by vFunct 1y ago
Yep. He has no idea what he's doing in media, or much else really (outside of AWS)
98.
by vFunct 1y ago
Sanction all companies and any VC fund that continue to do business with Israel.
99.
by vFunct 1y ago
Would this work with DaisyUI components?
100.
by vFunct 1y ago
How far behind Intel are they? Does Rapidus have backside power delivery?
101.
Intel to Lay Off 15% of Workers, Cancel Billions in Projects in Bid for Rebound(wsj.com)
8 points by vFunct 1y ago|2 comments
102.
by vFunct 1y ago
I'm surprised the US just doesn't fund a new fab company or consortium, like Japan did with Rapidus. But I guess "too much socialism"
103.
by vFunct 1y ago
Wafer-scale integration was done decades before.
104.
by vFunct 1y ago
You can if you use hyper spectral imaging...
105.
by vFunct 1y ago
Much faster than Claude Sonnet 4 with similar results.
106.
by vFunct 1y ago
Do we know if the full model is FP8 or FP16/BF16? The hugging face page says BF16: https://huggingface.co/Qwen/Qwen3-Coder-480B-A35B-Instruct So likely it needs 2x the memory.
107.
by vFunct 1y ago
Anyone else try this?
108.
by vFunct 1y ago
Why not both? LLM+Wikipedia RAG
109.
LLM Knows the Future: Uncovering Its Multi-Token Prediction Potential(arxiv.org)
24 points by vFunct 1y ago|0 comments
110.
by vFunct 1y ago
Basically it's only usable for rural or oceanic use. Impossible to use in a neighborhood of 10,000 people. And Elon wants the FAA to use Starlink to communicate between air traffic control centers, instead of the higher-bandwidth and m
111.
by vFunct 1y ago
Are non-US citizens banned from working on <2nm GAA technologies? What are the ITAR limits on non-US citizens from working on these tech? Can we incorporate in another country to allow hiring on non-US citizens to work on <2nm GAA tec
112.
by vFunct 1y ago
Any idea when we'll get a new protocol to replace HTTP/HTML for agents to use? An MCP for the web...
113.
by vFunct 1y ago
Pretty sure Apple is going to release WebGPU support in Safari in Mac OS X 26 Tahoe as well. There was a WebGPU video in one of the WWDC sessions.
114.
by vFunct 1y ago
Are people supposed to be against Hamas? Aren't Hamas heroes for fighting against the European invaders?
115.
by vFunct 1y ago
He describes 78,000 public pull requests per engineer over 53 days. LMAO. So it's likely 99.99% LLM written. Lots of good info in the post, surprised he was able to share so much publicly. I would have kept most of the business proces
116.
by vFunct 1y ago
YouTube.
117.
by vFunct 1y ago
You have to audit the risky tools because the system you are using was terribly designed. Again, it's the system's responsibility to make sure you don't fail, not your responsibility.
118.
by vFunct 1y ago
So if people aren't aware, you can have AI reimplement CUDA libraries for any hardware, as well as develop new ones. You wouldn't believe me if you didn't try it and see for yourself, so try it. NVidia's CUDA moat is no
119.
by vFunct 1y ago
This is my pain point in Django as well, the query DSL. 10+ years with Django and I still don't know how it's supposed to work, given it's all done within the namespace of a function call. Fortunately, I now have AI to write
120.
by vFunct 1y ago
Yah no that's not how it works. It's the system designer's responsibility to make sure you can run correct code without you reading it. If you have to read it, then your system has already failed.

More ›