Jabrov

7 ms·

1.
by Jabrov 1mo ago
It might still be useful for classification usecases
2.
by Jabrov 1mo ago
My eyes glazed over the AI slop that is this article. What are you even trying to say?
3.
by Jabrov 1mo ago
Is there any actual evidence that they did?
4.
by Jabrov 2mo ago
I’m so sick and tired of AI slop. I spend all day at my desk talking to LLMs for work. Then I come on here to escape a little and see what people are working on, and it’s just more LLM text. Good for you that you built a game engine or a ga
5.
by Jabrov 2mo ago
You need a static IP address for this to work is the downside, and depending on where you live and who your provider it it can be difficult and/or expensive.
6.
by Jabrov 2mo ago
Yes: cost, speed, and reliability. But all of those things are improving at shocking speeds, so I think we’re on a path where code is losing value quickly.
7.
by Jabrov 3mo ago
The letter is more about privacy, not freedom. They're arguing X is a massive privacy risk and should not get any exemptions.
8.
by Jabrov 3mo ago
It's not about being pointless, it's just plain wrong. The median (not average) household income in the US is 80K USD. p25 is 40K. p10 is 20K. They're struggling, sure. But I wouldn't call that abject poverty.
9.
by Jabrov 3mo ago
Who? What did he want?
10.
by Jabrov 3mo ago
The codes can trivially be shared in this case
11.
by Jabrov 3mo ago
Same
12.
by Jabrov 3mo ago
My eyes just glaze over when I see every file being listed and its purpose. Just because you CAN do that with LLMs doesn’t mean you should. What are you really trying to communicate?
13.
by Jabrov 3mo ago
Sooo it’s just streaming a game to your browser. What does this have to do with Dwarf Fortress? There’s tons of generic solutions like this that would work the same way for DF as well. It’s giving AI slop.
14.
by Jabrov 3mo ago
Transformers are superior
15.
by Jabrov 3mo ago
They've lost the plot, especially with the spa. And a billion scans a month is absurd. Is this some AI hallucination post?
16.
by Jabrov 3mo ago
They almost certainly did filter, but there’s always false negatives with this kind of stuff
17.
by Jabrov 3mo ago
Holy crap. Had to pick my jaw up off the floor. I hope you get some kind of acknowledgement or bounty for this. Kudos for having the willpower to resist sending a message to millions of people and sparking a global phenomenon!
18.
by Jabrov 3mo ago
Have you ever tried that? For me, the lag was the tiniest bit perceptible and was enough to make me sick
19.
by Jabrov 3mo ago
A million AI researcher voices at big tech companies suddenly cried out in terror and were suddenly silenced
20.
by Jabrov 4mo ago
Stop using ollama
21.
by Jabrov 4mo ago
It's yet another surface for dependency attacks
22.
by Jabrov 4mo ago
Why would there be less gridlock if people were in a driverless car instead of a regular car?
23.
by Jabrov 4mo ago
"The (annoying) solution is to package our server up into a connector/plugin and release it for each and every LLM client out there." Isn't this literally against the entire point of MCP
24.
by Jabrov 4mo ago
Hide some of your ballooning AI spend behind your ballooning AWS bill
25.
by Jabrov 4mo ago
I think you might have misunderstood the headline and maybe not read the article, because that's not really relevant to what's being announced here
26.
by Jabrov 4mo ago
I’m not American, and reminding OP that we’re not all American on this site
27.
by Jabrov 4mo ago
“Our cloud services are inundated” … says company that killed product from working offline and forced it to be connected
28.
by Jabrov 5mo ago
They absolutely are. The “maximum context window” of a model is a byproduct of the context length it was trained on. If your model only ever sees 8K token samples during training, it won’t be as good at 128K context length than if you had t
29.
by Jabrov 5mo ago
You are not we
30.
by Jabrov 5mo ago
I think part of the problem is that a lot of designers are simply not that good at what they do

More ›