thelastparadise

7 ms·

1.
by thelastparadise 2y ago
I had two, then only one once the liquor store down road closed up shop.
2.
by thelastparadise 2y ago
But do we know that the same techniques won't scale if trained in the huge clusters?
3.
by thelastparadise 2y ago
The megga disappointment is o1 is performing worse than o1-preview [1], and claude 3.6 had already nearly caught up to o1-preview. 1. https://x.com/nrehiew_/status/1864763064374976928
4.
by thelastparadise 2y ago
I could a car for that kind of money!
5.
by thelastparadise 2y ago
Is it working?
6.
by thelastparadise 2y ago
Do they denied their surgery or did they get it but bankrupt after?
7.
by thelastparadise 2y ago
What do you should be done with him then?
8.
by thelastparadise 2y ago
So what you're saying is Intel, or any other would-be NVIDIA competitor, needs to put out fast interconnects, not just compute cards. This is true. I'm not sure your argument stands when it comes to OP's idea of a single card
9.
by thelastparadise 2y ago
Thank you. No one ever talks about just how easy it is to add GPIO to anything.
10.
by thelastparadise 2y ago
Yeah, this is is true. I've been taking for 3 months now --best consistent sleep of my life, but definitely occasionally the most terrifying sleep paralysis I've ever experienced, or even heard of. I know it sounds ridiculous, but
11.
by thelastparadise 2y ago
If they did that to me I would take it as a sign of disrespect and resign immediately. Oops now you gotta hire a new one.
12.
by thelastparadise 2y ago
Several cases like this on here. These would be valid wrongful termination lawsuits.
13.
by thelastparadise 2y ago
Lol. Wonder how often toxic teams end up PIPing the actually highest performing people on the team, leading to a salt lake effect and slowly dieing corporation.
14.
by thelastparadise 2y ago
Your point about boilerplate is key, and it’s why I think MCP could work well despite some of the concerns raised. Right now, so many of us are writing redundant integrations or reinventing the same abstractions for tool usage and context m
15.
by thelastparadise 2y ago
That is done with multiwan in opnsense or mwan3 in openwrt.
16.
by thelastparadise 2y ago
So no computer use (pixel-level understanding). That's disappointing as the devtools approach always has limitations. Kura agents, Runner H, and scrapybara will all end up more reliable than you.
17.
by thelastparadise 2y ago
YouTube no longer has a monopoly on internet video distribution. If you notice, all the major platforms now have good support for direct posting of videos. Further, it's now easier than ever to decouple one's online life from goog
18.
by thelastparadise 2y ago
Likewise. Got an error at first, then it was working fine.
19.
by thelastparadise 2y ago
This can't be real.
20.
by thelastparadise 2y ago
This kind of thing is why I got into engineering in the first place. There's so much more to it than money.
21.
by thelastparadise 2y ago
not working
22.
by thelastparadise 2y ago
What do you think will happen that is worth relocating your entire family twice?
23.
by thelastparadise 2y ago
Java does this way better.
24.
by thelastparadise 2y ago
Watson did it too, a while back.
25.
by thelastparadise 2y ago
You can trust it if you configure it correctly.
26.
by thelastparadise 2y ago
The NPCs need a model of the world in their brain in order to act normal.
27.
by thelastparadise 2y ago
I wonder how LLM performance is on the higher core counts? With recent DDR generations and many core CPUs, perhaps CPUs will give GPUs a run for their money.
28.
by thelastparadise 2y ago
At least not until LLM gains hit a wall. So far every open weight model has far surpassed the previous releases at the same model size.
29.
by thelastparadise 2y ago
Orders of magnitude slower.
30.
by thelastparadise 2y ago
> This is the end of the Internet, goodbye. Beg to differ. We have to consider the fact that human prompting + selection of outputs is essentially RLHF, so the models can and will continue to get better over time. It's not the end o

More ›