Hacker Newsnew | past | comments | ask | show | jobs | submit | xchip's commentslogin

Well if you are shy, you have two options. Support going back to the office so everybody commutes there and you can easily talk to people without putting much effort... Or take it as a growth opportunity for yourself, take advantage of the extra time that you have saved in that useless commuting, and try to talk to people outside work and make friends as everybody else does.

Please believe in yourself!


Yeah, em-bee. You should just be not shy.

As a fellow shy person, I feel the pain when someone basically just tells you to get over it.


you didn't understand or even read what i wrote. my point was that talking to people in settings where talking is required is easier. and that is NOT the office, unless we are actually working on something together, but then i can also work together remotely, which works just as well.

try to talk to people outside work

you mean like what i wrote?:

working from home doesn't make things much worse. but, it allows me to avoid social isolation through other means. the advantage of going out to seek friends is that you can choose where to go, and you can go to places that are more open to interaction than the people at work


Nvidia fully supports remote work, people love it and they are doing great, not sure why others companies have issues with this.

NVIDIA is hiring wherever the talent is, learn from NVIDIA, stop whining.

It's not personal as it requires "an API key for at least one of: OpenAI, Anthropic, OpenRouter"


This article could have been summarized in three paragraphs.

I'm really hating this trend of diluting content by giving useless testimonials, random anecdotes and delaying the resolution of the subject as much as possible.


People like reading.

You could summarise all of Ender's Game in a couple of sentences but, guess what, that wouldn't be particularly pleasurable.

Not everything has to by hyper-efficient. More importantly, not everything has to be tailored specifically for you. It's OK that other people like reading long-form content.


Here, have a summary of In Search of Lost Time, courtesy of Wikipedia: The novel recounts the experiences of the Narrator (who is never definitively named) while he is growing up, learning about art, participating in society, and falling in love.

Hope you enjoyed it, no need to thank me.


You're REALLY going to hate this one weird trick Jesus did with all of the Old Testament, then. (and preemptively, with the New Testament, too!)


I also receive too much spam, I'll believe in their AI whenever they are able to fix spam.


spammers also use AI


There are about 150.7 million active iPhones in thr US

So, you are getting $1.66 back.


People who submit qualifying claims can receive $25 for each eligible device, “which may decrease or increase up to $95 per device, depending on claim volume and other factors,”


It’s for the iPhone 15 pro and 16 lineup


This is the general idea of how I understood it works.

- It's based on WAN a video model, it generates a frame at a time in latent space that we later we decode into RGB

- We keep the latent space and its RGB of each frame in a database. Along with the RGB we compute depth so we get a cloud point (RGBD). This will be used for persistence.

- For each new frame we check which past frames have their point cloud contained in the camera frustum. We take the top 3 frames with more overlap and get their latent space.

- We feed these 3 frames to WAN via the cross attention layers for conditioning and that is how we achieve consistency

- The RGBD can be used to generate a gaussian splatting of the scene.


"This transforms a vector of arbitrary real numbers into values between 0 and 1 that sum to 1"

Not really, softmax transforms logits (logariths of probabilities) into probabilities.

Probabilities → logits → back again.

Start with p = [0.6, 0.3, 0.1]. Logits = log(p) = [-0.51, -1.20, -2.30]. Softmax(logits) = original p.

NN prefer to output logits because they are linear and go from -inf to +inf.


Softmax is defined over an arbitrary vector of raw real numbers. Stating that those inputs are "logits" is applying post-hoc semantics to what the model is learning. One of the key properties of a softmax is scale invariance, (e.g. softmax([-1, 1, 3, 5]) == softmax([9, 11, 13, 15])) and so it is easiest to just think of it as operating on a vector of unnormalized raw scores, which is the more colloquial definition of logit.

(also, log(p) is not the formal definition of a logit)


(meant to say, scale-invariance of probability ratios, or shift-invariance of the inputs)


It's still true that softmax transforms arbitrary vectors into probability vectors.

In your example you'll also get the original `p` with just `exp(logits)`. Softmax normalizes the output to sum to one, so it can output a probability vector even if the input is _not_ simply `log(p)`.


Logit is log odds, not log probability: logit(p) = log(p / 1 - p)


Feels like a bailout. Belgium’s playbook is simple: skip maintenance, let it decay, then replace it on the taxpayer’s dime.


No, it was a deliberate strategy from the green party that were very anti-nuclear. They replaced it with gas power plants. And if you think this is a joke, no it isn't.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: