NostraDavid
[features.multi_agent_v2]
enabled = true
wait_agent_enabled = true
min_wait_timeout_ms = 10000 # 10s
default_wait_timeout_ms = 300000 # 5m
max_wait_timeout_ms = 3600000 # 1h
Seems to do the job and reduce usage; I just ran Astra for ~5 hours (using a goal) and it used the last 30% of my usage. And now they released GPT-6 Sol and Luna (which is basically 5.6 Sol and Luna, but a bit better and also 50% cheaper) ;_;Taalas (Acquired by AMD, back in August) created Jimmy[0], a little chat app that runs on a POC chip with ~14k tps. Yes, 14,000 tokens per second. Sure, it's just a 8B model or so (Llama 3.1 8B), but I can imagine that having a 1.58-bit model might be helpful for their next chip.
Heck, what would happen if you used a dLLM (d for diffusion)?
I'm hoping this will do better than my setup.
[1]: using the BlenderMCP + a bunch of skills I found online
That being said: I can imagine that IF any of them generates the design of a new drug that cures _any_ cancer right now, it will still take several years before seeing results - creating the drug must be feasible, then it must be lab-tested on animals, then on humans, get approval, etc, etc.
So I'm crossing my fingers on this one, but I won't hold my breath because 5+ years is a long time.
Is that Hezbollah's little brother?
Are the available compression levels not 1-22? Though that might depend on the specific library used - the official lib at least uses 1-22: https://github.com/facebook/zstd/blob/dev/programs/zstd.1.md...
Turns out we can squeeze quite a bit more compression performance out of DEFLATE - ~10x in certain instances, 2x as a base minimum (read the paper for details).
Why would I care if I don't notice? I'm paying them to fix my literal plumbing, not to proverbially suck my dick. If they do the job I pay them for, and they're not giving me shit for who I am, why would I want to get them fired?
This uni-context feels like a very silly idea in practice.
I want to be able to discuss taboo ideas in private, without getting globally cancelled for something I that might be discussed out of my mind. No thanks to the uni-context.
context-bound > uni-context, for at least the Germanic-speaking world.
Not the concept itself, but the insane numbers. Even South Africa is having "anti-migrant" protests (by the _black_ population; important detail, due to history).
Having 1-2% of your population come in as migrants* is pretty nuts; no negative migration afterwards; number only goes up. I cannot see how this is going to end well in the long run.
*: This is for the Netherlands, for the last 5 years since 2024 (that's the latest numbers I got from our Central Bureau of Statistics (CBS)). That is _just_ economic migration. It's insane. I made some visualizations: https://tbataafschebroederschap.nl/projects/autochtoonse-ned...
http://apestronauts.com exists since 2012, so that seems a decent potential.
Found it by throwing your comment into ChatGPT.
Then call it "Digital Sovereignty" instead. That's what we call it in the Netherlands.
> but it’s the term the industry is using
OK, and? Sounds like it is time to push back. It's like talking about "lower" and "higher" education. Just call it practical and theoretical education instead. Much more descriptive; doesn't talk down to people we seriously need in the coming decades.
Only if you're American.
The only thing I miss is that I want it to keep an OPML file up-to-date so I can autosync it to a repo, but it doesn't, so I have to export the OPML file sporadically. It's acceptable.
And then companies try to push for more Halal meat, because there are fewer rules to account for, when it comes to Halal (great way for them to skirt the law, legally). The chicken supposedly tastes better, though.
Also using Firefox with "Youtube-shorts block", which turns a short into a normal video, so I can't start doom-scrolling.
It cut procrastination-times down enough where I now feel I can program again.