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

>valid proof would never become invalid proof

Somebody can give their proof of age to another person.


And? Presentation of someone else's valid credentials is not fixable by any privacy-preserving mechanism. You can set an expiration date in order to rotate them, and they can be fast-rotating.

In any case, it's a moot point: the correct amount of required identification is zero.


> Presentation of someone else's valid credentials is not fixable by any privacy-preserving mechanism.

And that is precisely why governments will never implement a privacy-preserving mechanism, which is exactly my point.

Compromised tokens would be trivially google-able within a day otherwise.


It's a CLI app, it's already configurable. Every good terminal emulator lets you set custom palettes.

>Every good terminal emulator lets you set custom palettes

Not differently for each program's output.


Which is a good reason to stick with the de-facto standard of red for bad and green for good.

unless you're colour blind

If you're color blind, you change the palette in your terminal emulator so "red" and "green" become different colors you can distinguish. It even works for rarer forms of color blindness. This works best when people follow the de-facto standard.

Red here does not mean #ff0000. it means color 1. in the 4 bit colors palette

> Red here does not mean #ff0000

For you maybe.

"Look it is not gray on gray is black #777777 on white #333333". /s


If for something unrelated to good/bad looks good, I'm using red. Ditto for green.

Sure, if it was a status indicator and I used red for "good", I can see the point. But over the last few years I've had too many people tell me "Don't use red, people will think something is wrong" for things not semantically tied to good/bad.

People wear red clothes. They buy red cars. They eat red food. They date red heads. Red is OK.


The standard terminal palette is only 16 colors. Even if you compress them all into the green-to-blue color range, it's still possible to distinguish all 16. The user can change "red" and "green" to whatever they like in the terminal preferences and then every 16-color app will be accessible with no additional effort from anybody.

The original post claimed they were "running hundreds of concurrent agents":

https://cursor.com/blog/scaling-agents


It was 2,000 concurrent agents at peak.

I'd still be surprised if that added up to "trillions" of tokens. A trillion is a very big number.


16 million a second across 2000 agents would be 8000 tokens per second per agent. This doesn't seem right to me.

I mean, its right there in their blog - https://cursor.com/blog/scaling-agents

"We've deployed trillions of tokens across these agents toward a single goal. The system isn't perfectly efficient, but it's far more effective than we expected."


During inference, most of the memory is read only.

Sounds fair. That's not the kind of machine I'd want as a development system though. And usually development systems are beefier than production systems. So curious how they'd solve that.

Yeah, it is quite specialized for inference. It's unlikely that you'd see this stuff outside of hardware specifically for that.

Development systems for AI inference tend to be smaller by necessity. A DGX Spark, Station, a single B300 node... you'd work on something like that before deploying to a larger cluster. There's just nothing bigger than what you'd actually deploy to.


HBF, like expensive HBM, is targeted at AI data centers.

  The KAIST professor discussed an HBF unit having a capacity of 512 GB and a 1.638 TBps bandwidth.
PCIe x8 GPU bandwidth is about 32GBbps, so HBF could be 50x PCIe bandwidth.

If you're practicing your serve alone it's convenient to have a lot of tennis balls. It's quicker to gather all the balls in a big batch than individually.

For me, learning that normal people go about their days constantly hallucinating had the opposite effect. I think it could partly explain some problems in society, e.g. people's susceptibility to advertising.

I think your implicitly getting at something here. Both are dealing with an inferiority/superiority dynamic. The suggestion of a group you identify as being less, causes a predictable reaction to characterize the other (non-aphantasia) as problematic/hallucinating (i.e. broken/lacking). This ties back to the post where the author speaks of feelings of inadequacy (shame, etc...) about being unable to visualize, again signs of an inferiority complex. While such complexes may be traced back to particular memories or events, they're also habits of thought which are common place and culturally reinforced, so much so that they seem quite normal. For example, the culture of idol worship, like raising up of tech heroes while implicitly lowering your own self worth, which happens often on this site.

The fact that the author doesn't mention the details of the memory or events of the day also suggests shame and concerns of being judged for them.

The good news is they are writing about their struggles which suggests their willingness to work with these fears.

I think the answer probably isn't about pretending you're not better or worse, but accepting that being better or worse at something doesn't change your inherent self worth. Accepting that your not in control of many of your conditions and conditioning can free the mind from a sense of guilt and the fear around judgement of yourself and others. Hopefully this helps the author and those who struggle with notions of identity and self worth.


> The suggestion of a group you identify as being less, causes a predictable reaction to characterize the other (non-aphantasia) as problematic/hallucinating (i.e. broken/lacking).

No, it's not a defensive/counterattacking reflex. The thought of people hallucinating all the time is terrifying to me, because hallucination is a sign of something being very wrong, like schizophrenia. After getting past the language barrier and finding out these were "mental hallucinations" rather than "visual hallucinations", it's slightly less scary, but still unsettling for me to think about. Finding out that visualization was actually a thing meant that idioms that I thought were metaphors or superstition were suddenly something the majority of the population takes literally. People who have "invisible friends" talking to them all day long scares me though.


Also, the metaphorical "little voice in the back of your head" that tells you what you're about to do is a bad idea.. Apparently people really hear that too? It would be nice to read about the differences between genuine hallucinations and mental imagery or sound from someone familiar with both. Obviously there are differences but people just get offended or simply weirded out most of the time when you ask.

Also you might find it interesting to read Jaynes' Breakdown of the Bicameral Mind. It's evidently controversial and not a viewpoint to uncritically adopt wholesale, but it does get you thinking about mental visualization/audiolization vs. hallucinations etc... and contains some intriguing historical anecdotes.


> It would be nice to read about the differences between genuine hallucinations and mental imagery or sound from someone familiar with both. Obviously there are differences but people just get offended or simply weirded out most of the time when you ask.

I thought the same, but after reading this I'm beginning to wonder whether or not there is actually a difference: http://www.sciencedaily.com/releases/2026/01/260122074033.ht...

It seems like the "voice in the head" is distinguished from real voices by a mechanism similar to how tickling yourself doesn't cause the same sensation as another doing the tickling. People with inner voices and visualizations might actually be hallucinating all the time, they're just aware of it and not being misled by their senses like a schizophrenic would be.


Very interesting! Also interesting how unquestioningly "thoughts" are equivocated with "internal voices" in both the press release and the paper.

Regardless of current mechanism, susceptibility to advertising would still be present even if all currently exploited cognitive pathways were removed or deactivated across all human minds, as the advertisers would keep experimenting until they found another one.

The easiest solution is to use the default font. This has the additional benefit of being the most legible font for every reader, because it's the one they have the most experience reading.


This is IMO the correct decision, and I say that as a regular user of middle-click paste. Middle-click paste is a classic convenience/security tradeoff. It's faster than using the keyboard, but it makes it very easy to accidentally leak confidential information. In all such tradeoffs, the default setting should favor security.

Only if the option was removed entirely would I have reason to complain.


In theory I'd agree as well.

In reality such features are often dropped, no settings in gnome-settings, and non default features are not tested in new builds.

So you end up googling how to get that feature back and the answer is manually navigating the config tree or using gnome-tweaks, both with large disclaimers that any changes might break your system.

No thanks, I'd rather have left click to select and middle click to paste, do we really have to involve the keyboard by default?


I half agree, but rather I'd say secrets shouldn't go into clipboard. Have some kind of separate channel for it. It is insane to me that so many workflows rely on putting secrets into a box any app or even webpage can peek into.

I love selecting some text in the terminal, then just middle-clicking and... aaahhh!


Classical labels were recording digitally even before CD players existed, to avoid the generation loss of recording to tape before transferring to vinyl. These recordings were later released on CD and mostly sound great.

Dire Straits - Brothers In Arms (1985 release, first CD to sell over a million copies) also sounds great, and IMO better than most modern releases.

Some early CDs were recorded using pre-emphasis, similar to the RIAA equalization used with vinyl records. CDs using this have a flag set in the metadata to tell the player to apply a matching de-emphasis filter. I sometimes see people blaming digital production for early CDs sounding "thin". I think it's more likely they heard rips of CDs using pre-emphasis that didn't have the proper de-emphasis applied.

An average CD from the 80s sounds better than an average CD from any other era, because it pre-dates the loudness war, and because it's intended to be played on a good home stereo (which if you were buying CDs back then you could probably afford).


I was thinking about back catalogs, not necessarily new recordings. Things recorded pre-1980.


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

Search: