pants2

32 ms·

691.
by pants2 2y ago
No but I'll think about writing one up!
692.
by pants2 2y ago
This will tie nicely into my favorite way to deploy services these days: 1. Use PyInfra to set up Docker and Tailscale on remote hosts and any other setup. Open the Docker port to your Tailnet. 2. Use the Docker provider for Terraform to se
693.
by pants2 2y ago
I made a game like this the other day just for fun! It worked pretty well but it's more of a novelty and not really fun to whistle for more than 30s.
694.
Global Plastics Treaty(theoceancleanup.com)
4 points by pants2 2y ago|0 comments
695.
by pants2 2y ago
And who can forget their famous headline "Jurisprudence Fetishist Gets Off On Technicality"
696.
by pants2 2y ago
This is the best birthday present I could have asked for. The Onion News Network has the funniest, most prescient videos online and I love going back to watch them. Even if the Onion comes back half as funny as they were it's going to
697.
by pants2 3y ago
I recently upgraded from Windows 11 to Windows 10 and I love it - it feels more user friendly and customizable! For instance I can choose not to group my applications in the task bar. There's also no intrusive ads, and I don't nee
698.
by pants2 3y ago
At least here in SoCal it's not uncommon for the prevailing speed on the highway to be 20+ mph over the limit. I've definitely passed cops doing 85 in a 65 and have never been pulled over for speeding. Seems like it's time to
699.
by pants2 3y ago
I'd be interested to hear how Llama 8B with long chain-of-thought prompts compares to GPT-4 one-shot prompts for real-world tasks. In classification for example, you could ask Llama 8B to reason through each possibility, rank them, rat
700.
by pants2 3y ago
There are dozens of AI chip startups out there with wild claims about speed. Groq seems like the first to actually prove it by launching a product. I hope they spur a speed war with other chipmakers to make the fastest inference engine.
701.
by pants2 3y ago
Another good resource: https://artificialanalysis.ai/
702.
by pants2 3y ago
It's approximately a 4 hour drive by car. The train might be able to do it in 2 hours, but of course there's the overhead you mention. The drive costs around $40 in gas one-way. If this train is indeed $80 per ticket one-way, then
703.
by pants2 3y ago
I have a custom app that does this. It's not open source unfortunately but you can make one in a couple hours with the Gmail API and something like Llama Index.
704.
by pants2 3y ago
Copilot money seems pretty decent so far. You might also just embrace the spreadsheet and use Tiller to scrape your info into one. At one point I tried to run my finances like a business and use QuickBooks, generating quarterly statements o
705.
by pants2 3y ago
My dad was as programmer and tried hard to get me into coding early, but according to him I resisted and refused every time. Then I took one programming class in college and realized I was really good at it, so something must have stuck fro
706.
by pants2 3y ago
Yes, this is an awesome feature on Android! > Android Oreo came with a new "smart text selection" feature that uses AI code to make smart decisions about selecting text such as addresses and phone numbers.
707.
by pants2 3y ago
At least with Mac we have Raycast as a vastly superior alternative to spotlight! You can also install GBoard as an alternative keyboard on iPhone and use it's superior autocorrect. But sadly there is no way to replace Siri.
708.
by pants2 3y ago
I'm surprised to hear there are fewer than 4,000 Cybertrucks because there are half a dozen in my small neighborhood and I see them driving around everywhere in SoCal!
709.
by pants2 3y ago
This doesn't even mention the most obvious one to me: Google Assistant. It's way, way better than Siri and well integrated (at least on pixel phones that I've used).
710.
by pants2 3y ago
LLMs are excellent at presenting more common diagnoses first. With Google you can easily find some rare and scary disease that fits your symptoms, that trips up a lot of casual Googlers.
711.
by pants2 3y ago
Considering that textbooks are probably the single highest quality source of training data for LLMs, I would be very surprised if OpenAI wasn't buying and scanning textbooks for their own training data (including books that aren't
712.
by pants2 3y ago
In theory, what's the best way to take out a robot like Atlas (or next year's more advanced military model)? It seems like they could be made electromagnetically shielded, waterproof, bulletproof, etc. Maybe just armor piercing ro
713.
by pants2 3y ago
Can't wait for 8x22B to make it to Groq! Having an LLM at near GPT-4 performance with Groq speed would be incredible, especially for real-time voice chat.
714.
by pants2 3y ago
Ollama runs locally. What's the best option for calling the new Mixtral model on someone else's server programmatically?
715.
by pants2 3y ago
Sounds like you might enjoy QuestDB. It's Postgres-compatible but oriented toward time series and streaming data. It's easy to install and play around with and comes with a built-in explorer UI.
716.
by pants2 3y ago
This is what Comic Con does but it's more like 130,000 people
717.
by pants2 3y ago
It's probably also worth checking your ping to these servers. For me, Cloudflare and Quad9 are around 10ms, while Mullvad is around 40ms, so I take that into consideration when choosing a DNS provider.
718.
by pants2 3y ago
I assume that's https://en.m.wikipedia.org/wiki/Andrews_%26_Arnold The minute they come to America I'm switching.
719.
by pants2 3y ago
This is a great idea. Apparently some people have reverse-engineered the network authentication to allow you to bypass the AT&T modem, but it looks pretty hairy.
720.
by pants2 3y ago
Mullvad has a public DNS service that I'd trust.

More ›