Hacker News Reader

fauxpause_on Bakhmut Is Gone: An Aerial Look at the War’s Destruction

I mean it’s sucks that this guy died but it was still a successful op. 9 months of Russian offensive and they got nothing of value.

fauxpause_on Bakhmut Is Gone: An Aerial Look at the War’s Destruction

It wasn’t particularly clever. Russian command was just particularly dumb.

The last Ukrainian counter offensive reclaimed vast swaths of lands. It’s not unreasonable to think they could get in range of shelling Crimea.

fauxpause_on Bakhmut Is Gone: An Aerial Look at the War’s Destruction

Yep. They’ve been very transparent about it.

fauxpause_on Bakhmut Is Gone: An Aerial Look at the War’s Destruction

Ukrainian military and third party analysts have been very clear on this for a long time. Bakhmut had little strategic value. But Russia was hellbent on taking it. So Ukraine was using it as a way to pin down Russian forces and kill them en masse as they used their stupid human wave tactics. It’s worked well.

Russia spent 9 months on this offensive. They got Bakhmut and lost immense combat power. Now Ukraine is readying a counter offensive.

fauxpause_on Bakhmut Is Gone: An Aerial Look at the War’s Destruction

I suspect it will be within full Ukrainian control within less than a year. It’s not a strategic priority. And it is costly to try and take inch by inch. So Ukraine is not going to be in any rush to take it back. But I also expect the priorities for Russian troops will be stretched extremely thin. Wagner appears eager to leave and blame the Russian MoD for failing to hold it.

The news today of Russian freedom fighters “liberating” a Russian settlement around Belgorod is the most interesting thing to happen in a while. Keep an eye on that one.

fauxpause_on Meta fined $1.3B over data transfers to U.S.

You’re right in that it ought to be compared to the scale of profits, not a percentage, as many run on a loss during growth. But profit is still what matters. Including the promise of future profits.

Talking of the future doesn’t help much because both numbers will change. And punishing a company based on its future state is… not possible

fauxpause_on Meta fined $1.3B over data transfers to U.S.

The accounting is not what matters. What matters is using your brain to to figure out if a fine is actually meaningful.

Comparing to revenue is a stupid way to think about things. Profit is the incentive to conduct business. Not revenue. And not global profit, but in this case Ireland/EU profits only, because that is the location fining them.

People are so eager. Every. Single. Time. To say that a fine does not matter even if it clearly outpaces multiple years of profits for the area given.

fauxpause_on Don Knuth plays with ChatGPT

I’m getting pretty titled at the number of people who are ignoring everything I’m posting and claiming temperature is some unknowable thing because Knuth does not know what it is. Look at my link. This is not a concept specific to them. It’s a single term in the softmax selection.

There is no reason to assume that OpenAI has changed the definition of this term.

fauxpause_on Meta fined $1.3B over data transfers to U.S.

They should be written as a % of profit from the area generating the fine

fauxpause_on Don Knuth plays with ChatGPT

My dude it’s not my fault if you don’t understand the concept of asking follow up questions for clarification. This isn’t like a Google search. The way you retrieve knowledge is different

fauxpause_on Don Knuth plays with ChatGPT

Yes. I have posted both a very nice link and a complete explanation from chat gpt 3.5 itself. It’s honestly not that complicated, especially for someone who is supposed to have any sort of authoritative view in the field.

I do not feel it is appropriate for you to say you have looked it up if you don’t know what it is besides an API input that affects randomness.

fauxpause_on Don Knuth plays with ChatGPT

Seems like a good answer. If you want a deeper answer, you should ask a follow up question. Here is my follow up answer to “can you explain it at a deeper level”

> At a deeper level, the temperature parameter in generative AI models like GPT-3 affects the softmax function, which is used to convert the model's output into a probability distribution over the possible words or tokens.

> The softmax function takes the logits, which are the unnormalized output values of the model, and applies the exponential function to each value, normalizing them to produce probabilities. The formula for the softmax function is as follows:

> P(y_i) = exp(logit(y_i)) / sum(exp(logit(y_j)))

> Where:

> P(y_i) is the probability of the i-th token in the output sequence. logit(y_i) is the logit value (unnormalized score) of the i-th token. > The sum is taken over all possible tokens. > The temperature parameter comes into play when computing the exponential function. The temperature value, denoted as T, is applied as an exponent to the logits before performing the softmax. Mathematically, the modified softmax function with temperature is:

> P(y_i) = exp(logit(y_i) / T) / sum(exp(logit(y_j) / T))

> The temperature parameter T controls the steepness of the softmax function. A higher temperature value (e.g., 1.0) increases the exponentiation effect, making the probabilities more spread out and the distribution more uniform. This leads to a higher likelihood of selecting less probable tokens, resulting in more diverse and random output.

> Conversely, a lower temperature value (e.g., 0.5) reduces the exponentiation effect, making the probabilities sharper and more focused. The most probable tokens are emphasized, leading to more deterministic and conservative output.

> By adjusting the temperature, you can manipulate the balance between exploration and exploitation during text generation. Higher temperatures encourage the model to explore more possibilities, while lower temperatures bias it towards exploiting the most likely choices based on the training data.

> It's important to note that the choice of temperature is a design decision, and there is no fixed rule for determining the optimal value. It often requires experimentation and understanding the specific behavior of the model to achieve the desired output characteristics

fauxpause_on Don Knuth plays with ChatGPT

https://lukesalamone.github.io/posts/what-is-temperature/

> Well, no, making a point of highlighting the points of your ignorance when discussing something is good. Especially when you are a notable expert in the broad field being discussed.

I disagree. Stating “whatever that means” indicates dismissiveness, not a transparent lack of expertise. Also, you should know what it means if you’re an expert.

This quote implies to me that he is actually a beginner when it comes to this technology but is expecting to be treated like an expert whose experience generalizes

fauxpause_on Don Knuth plays with ChatGPT

I don’t really care if he knows it or not. Best case he’s virtue signaling ignorance.

fauxpause_on Don Knuth plays with ChatGPT

Temperature is a measure of capriciousness. How likely is the model to choose a token that is not “the most likely” next token.

It’s not a big ask to look this up. But even if you don’t, making a point to show that you don’t know it seems bad.

fauxpause_on Attention, Hollywood: De-Aging Isn’t Working, So Please Stop Using It

I would have liked more stills that show bad de aging

fauxpause_on Attention, Hollywood: De-Aging Isn’t Working, So Please Stop Using It

How dare they put military coolness into a top gun movie

fauxpause_on Attention, Hollywood: De-Aging Isn’t Working, So Please Stop Using It

How old do you think he is supposed to be?

fauxpause_on Russia’s Nukes Probably Don’t Work – Here’s Why

The west is very clear that they’re not willing to allow the precedent of small scale nuclear weapons use without consequence. If Russia does this, we will massively ramp up our involvement in the war.

Nuclear weapons will not actually make a huge difference. Neither on the battlefield nor on civilian targets.

fauxpause_on Russia’s Nukes Probably Don’t Work – Here’s Why

Kind of buried the lede I think. The end states that tactical nukes are probably fine. Those are the ones most likely to be used. Although I think very unlikely still

fauxpause_on Russia’s Nukes Probably Don’t Work – Here’s Why

Launching a single nuke that might not work is a high risk low reward gamble too. It seems very unlikely to come up

fauxpause_on Don Knuth plays with ChatGPT

Seems like a bad example. You didn’t give it something to apologize for and used it as an example of it not apologizing

fauxpause_on 12-Year-Old to Graduate from College with Five Degrees

That’s literally a point I made in reverse. It’s too much work. For a good, real degree. This is not the story of a boy genius thriving on raw intelligence. I suspect this is a story of his parents manufacturing a story by figuring out how to get these degrees with the minimal possible effort.

Nobody gets 5 degrees in unrelated things out of personal interest. Especially not at 12. Normal interested people may take additional classes in other areas, but you don’t get 5 degrees unless you optimize for getting 5 degrees. Again, purely based on time.

The kid also had to be homeschool for normal school stuff the entire time. I say because I got annoyed and dug up additional info.

fauxpause_on 12-Year-Old to Graduate from College with Five Degrees

Idk, the cynic in me wonders what he actually had to do. Because for a lot of degrees the answer is basically nothing.

5 degrees should not be possible imo. Not because you can’t master 5 subjects but because you shouldn’t have time to do 5 classes worth of work.

It feels manufactured. Talking about the kid doing archery and going to 23 countries makes me feel this even more.

fauxpause_on Social network ad targeting can “listen into” conversations using “live photos”

If you accept that as true then you also have to accept that your voice is hopelessly copyable and defense against that is futile. So it’s not really important.

fauxpause_on Social network ad targeting can “listen into” conversations using “live photos”

There are billions of ants

fauxpause_on To help new students adapt, some colleges are eliminating grades

Because the basic foundational knowledge of biology is useful to understanding the world you exist in.

fauxpause_on Legend of Zelda game sells 10M copies in three days

Majora’s Mask is such an incredibly ambitious title. The game loop is just super cool. The aesthetic is this super interesting depressed melancholic cheerfulness of people coping with their impending death. OoT’s ambition is of course noteworthy in its own right, but damn Majora’s is just so cool.

fauxpause_on To help new students adapt, some colleges are eliminating grades

Imagine wasting time in school teaching things like geometry, history, and science; when the student wants to be playing basketball

fauxpause_on [dead]

Brain wave mapping to text is unique from person to person. They’re also hard to read and tend to require concentrated focus on specific patterns of conveying text (more like consciously shuffling to be understood rather than the reverse). I think this is an unlikely development. What is more likely is you have a personal device like a phone that you have reading your own thoughts and that’s just some leaky thing that corps can listen to.