planetpluta

I’d say two big differences are 1) human vs machine (especially when you get to the scale of something like Delta airlines) and 2) you have a lot more power in the negotiations you described! Basing it on 5 years of purchases and historical data isn’t a negotiation—it’s a “my way or the highway”

This is an interesting way to think about it. I would argue that flat pricing wasn’t for efficiency but “fairness”.

And also point out that AI driven price discrimination isn’t anywhere close to negotiated. You’re stuck with the price the machine gives you, with little to no recourse, short of rewriting your entire digital life!

Curious about the 100 pushups program app — do you have a particular one you like that you can share?

Edit: Followed the github issue and found the link!

A lot of apps allow you to AirPlay to multiple devices at once — would be neat to put this on a bunch of iphones to simultaneously play music

> I just assumed Vanta (the company) was doing a Show HN today and was confused at first glance

Did the title of the post change? At first glance the Show HN is a toy wireshark program very far from any Trust Management and compliance

I think the biggest hurdle would be complying with the TOS. Imagine that OpenAI etc would not be a fan of sharing quotas across individuals in this way

Anecdotally, I completely agree that consumption of media has shifted since the pandemic and could reasonable explain this gap

In a sense, the macro trend is irrelevant and mainly used as a talking point by the media.

Individuals experience the world from an individual level — it is easy to go along with any trend that fits your desired narrative because until it is at odds with your individual experience, it doesn’t really matter.

(I’m being a bit reductive and haven’t fully fleshed out this thought, but think the sentiment is accurate)

> Essentially we needed to teach LLMs (which are paid by the token) to only generate a small number of high quality comments.

The solution of filtering after the comment is generated doesn’t seem to address the “paid by the token” piece.

Something to consider is that it really could take 20+ years to train like a brain. But once you’ve trained it, you can replicate at ~0 cost, unlike a brain.

> This is deeply wrong

I can’t tell if Elon is saying that the tweet is misinformed or OpenAI’s behavior is reprehensible

Maybe not “immediate” but withholding rewards results in fewer researchers participating in bounty programs which defeats the purpose.

Okay that makes sense. Agreed, wouldn’t quite consider that “supporting” as far as I’m concerned.

I’m confused by how pushing to new branches would work on GitHub (or is the point that it doesn’t…)? Are you able to change the branch of a PR from `v1` to `v2` without making a new PR?

This tool looks great! I’ve been using difftastic lately, which does a fairly good job but struggles with big json files.

One feature I’ve yet to see is applying jq query syntax to the jsons before the diff

We don’t know the counter factual here… maybe if he called it “Unsafe Superintelligence Inc” they would have raised 5x! (though I have doubts about that)

I appreciate being able to quickly skim an article rather than watch a video - especially when in public without headphones!

Gary Marcus has been a firm advocate that GPT-5 is failing for a while now, based simply on the fact of them not having released it yet.

If/when OpenAI releases it, looking forward to how he reconciles his views.

> Take, for example, "Day," which is even, and "Say," which is odd. Since the "-ay" suffix is consistent, it indicates that "S" is the first letter in the target word.

Am I missing something? Given that information, couldn’t the D+a or D+y be correct?

(I understand those cannot be correct because of “May”, but not from the consistent suffix alone)

> there’s mostly no news after hours

As I understand it, most company earnings are released after hours and many important press releases are held until the market has closed to allow the news to be “digested” offline

Yes there is an `—-override` option you can use to specify the language in which a file should be parsed.

https://github.com/Wilfred/difftastic/blob/master/CHANGELOG....

Maybe their work has a policy against it, but surprising they never used their phone to take a photo either!

Agreed — isn’t the problem that nothing was actually being set aside for the insurance fund, not that the amount being set aside was based on a random number?

If a banana costs ~60¢ (depending where you get it!) and GPT-4 is 6¢ per 1k tokens… you’re better off using a monkey for everything > 10k tokens!

I think the rest of that sentence tries to explain the dismissal:

>> There are plants that don’t appear green, like the copper beech, because they contain pigments like carotenoids. But those pigments are not photosynthetic: They typically protect the plants like sunscreen, buffering against slow changes in their light exposure

The reason for non-green is not related to photosynthesis but some other process