One interesting thing that is relevant today -- I generally write drafts of READMEs in whatever model I am developing with, and end up rewriting almost all of it. Today I pasted my readme from claude into Gemini 4 Argon, and asked it to rewrite for natural reading and clarity, and got the current README. I'm impressed!
>These permission prompts are a part of a macOS subsystem called Transparency, Consent, and Control (TCC), although Apple doesn’t seem to use this name anymore.
I have a PiKVM on my Mac Mini for this reason, and I bet the author would be well served by one as well.
HDMI out works great on M1 and M2 chips, and DP alt mode output is nearly ready, and works if you compile an experimental kernel. I daily drive an Asahi Linux ALARM M1 Pro macbook, and I have almost zero complaints.
The Asahi installer does this automatically, it guides you through sizing. You run it in a macOS terminal. By default, Asahi will be the default boot option, and to get to macOS you hold the power button on boot until it says "Loading startup options..."
I have two M1 pro asahi macbooks, and I only use macOS for VirtualDJ or AirPlay.
Unless you run overnight, you could schedule a cron job to send a basic claude -p prompt such as "reply with hello" using haiku to align your usage windows. That's what I do.
Yes something like that is what I did, so I had my 5 hour reset window to be at 2 hours so I could work. But anthropic reset it so it went back to 5 hours.
I've been using the 35B-A3B today for some web scraping work, and it has been on par with Qwen3.8 27B at a much higher speed and at a higher quant (q4 vs q8). I'm impressed.
What you say is in agreement with someone I follow on X. My family has been traveling for two weeks and I have been looking forward to trying qwen3.8-37b when I get home.
Now I might try Ornith-1.5-35B first even though I don’t see an official MLX version.
I had to go down to UD-Q3_K_XL for Qwen 3.8 27B to get it to fit in VRAM and be usable, but I worry I'm gutting its intelligence somewhat. I too am interested in faster + more-usable alternative that can exchange blows with the Q3-dumbed 27B.
They may have utility in trying to look at the whole landscape of models, but are very misleading when it comes to making 1:1 comparisons or in developing confidence at to how a given model will deliver on your workflow.
Only relevant benchmarks are those you make yourself, targeted specifically for your workflows. Anything else is just number go up on a pretty graph, and every model out there is probably benchmaxxed to hell on the public ones anyway. Keep yours private.
These models have gotten a fair amount of attention -- we're hoping it's enough to get them added to some reliable inference providers and OpenRouter, at which point we'll run them on our full benchmark suite.
Effectively non-existent in Canada. Kijiji nominally exists, but I dual list everything I'm trying to get rid of on both kijiji/facebook out of principle and get like 20x more interest on facebook.
I found more that weren't accepted than I did that were in the couple minutes I played -- Michelle Obama, Melania Trump, Kate Middleton, Rosa Parks, Marie Curie, Ada Lovelace
Yes, there's quite a lot of names that somehow don't qualify: Margaret Thatcher, Sacagawea, Joan of Arc, Marie Antoinette, yet Madonna and Janet Jackson works just fine.. its really more of a "do you know what pop-culture person I'm thinking of" game.
reply