Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I have Claude max plan and the vscode claude dashboard plugin has logged about $4k worth of tokens in the past 2 months. I upgraded because I was using my weekly basic plan tokens in like 5 hours.

Likewise, I don't understand how anyone survives on the basic plans. It's weird seeing these two camps not understanding what the other is doing :)


Oh jeez, I wonder who on Earth could be behind this mass push for surveillance?

0. For the love of the game

1. Additional freedom for used purchasers, gifted purchasers, purchasers from a past era, whatever


so your company runs on a vibe-coded saas app, that sure is a confidence booster for your would-be customers.

Not sure what to make if your benchmark because GPT 5.5(low) ranks higher than GPT 5.5 (medium) -- #4 vs #9

Agreed, I just wanted to point out that it is still around!

I just checked 2007 and 1999 reports [1] a a bit and doesn't seem like they made such obvious warnings at those times.

I don't know much about economy and I just did some ctrl + f skimming, but this new 2026 warning is obviously more clear to me.

[1] https://www.bis.org/annualeconomicreports/index.htm?annualec...


They were limited by the tech of the time too. A modern clean sheet redesign of a 747/A380 class airplane would look very different given modern composites and improvements in engine bypass and compression ratio.

So the pitchforks will be the poors forming a mutual fund and busting/pump dumping the riches stocks?

I was also a fan of Hex, but development has stalled. I tried FluidVoice (found it buggy) and Handy (had problems with hotkey) before eventually landing on TypeWhisper[0].

It works very well, and has a few more bells and whistles without feeling bloated.

[0] https://www.typewhisper.com


What sort of measures would those be? The government also banned alcohol on campus (except Louisiana) but as a 20 year old I was too stubborn to care.

>What's the population of the suburb you live in? About that?

This comment shows you're being disingenuous. If a war killed 4 people, you'd be asking how big my family was. You can obviously understand the significance of that figure and it's nothing to do with suburbs.


Yeah, I forgot about these :)

This comment sounds like it's written by someone who doesn't interact with real people very often

Where is the weird part?

Kids need private spaces. If we overprotect them in the real world, and overprotect them in the digital world children will fail to explore all sorts of things.

The current US administration has gone a long way towards handing over leadership in AI to China.

> have to manually run the code in your head

If you're doing a good job, you have to do that anyway, or at least have enough of a spidey sense for broken code to know when to investigate and add an extra test case.

Something like 30% of the time at $WORK, interviewers report the candidate as having solved the problem when a closer inspection reveals UB, memory corruption, and other bullshit. The test cases pass, and I think that's part of the problem. You can't tune out and avoid deeply understanding the submission.


I know someone who went to Reed College, which has semi-famously not suffered from grade inflation[1]. They send your transcripts out with an explanatory note, so that the recipient will not view the graduate poorly when they see the numbers.

Interestingly, at Reed, there is a low emphasis (or even anti-emphasis) on grades — a student has to go out of their way to obtain them. Instead, emphasis is on written feedback and discussion, to understand one's performance on assignments.

All this to say: de-emphasizing grades in school is not necessarily a bad thing, and does not necessarily harm the reputation of the university. It can be a sign of good priorities (eg: learning, rather than numbers-gaming).

[1] https://www.gradeinflation.com/Reed.html


In my day job we tried creating a credit assessor tool using LLM as the credit assessor.

It did great, generated a report on the assessed business that was incredibly detailed and plausible.

Then I started running tests and getting into the details, and found that if you ran the same report on the same data, it generated completely different, still very plausible, results. I could run the same source data through the assessment process 10 times and get 10 very different results. We had to can the project and go a different route.

LLMs are designed to produce plausible results, not factual results. We can fix this when using them for software dev by using linters and tests (though we've all had the experience where the LLM invents an API endpoint). I would not trust raw LLM output in any situation where that kind of testing and verification capability isn't present.


> they scale out in 2D, we scale up in 3D.

This actually helps a lot, thanks.

> Instead of spreading SRAM across a wafer, we stack DRAM on top of the logic

Is this done with current manufacturing technologies? Does it require a special process?

> no streaming, no off-chip memory at all. ~1 kW, not 23 kW

Is this for an individual compute unit? Compared to Cerebras, what's the ratio of power used vs compute?


I dunno, I could see it working.

I do something similar with reviewing code: I have one agent write the code and another reviews it, then they go back and forth for a bit improving the code. Seems to yield better results than one agent alone.

Seems like a similar principle.


I suspect you are being downvoted because your answer is AI-generated, but I found it very clear and will upvote.

Yeah, but it's everything getting fuzzed, including the parts you care about.

Of course not. Education is free on the internet. University provides motivation and credentials. The motivation part is important because most people can't stick to such rigorous education for so long without some external force. Universities need to enforce rules against cheating as part of the motivation service they're providing.

We're paying more now than we paid in the late 2010s, but less than we paid in the early 2010s.

The ore copywriter world did not have any means of mass duplication.

I'm going to lose interest in this conversation soon, but to clarify for anyone else who might be reading: wearing body armor is not common, and tends to be evidence that you are preparing for violence. When you commit a crime of violence, such as attempted murder, preparations ahead of time become evidence of prior intent, which raises the level of the felony involved. E.g., if you get into a verbal confrontation and, in the heat of the moment, pull out a gun and shoot the guy you were fighting with, that's murder, but it's not premeditated murder. It'll be murder in the third degree or second degree, depending on circumstances (and, I think, the definition of degree might vary from state to state).

But if you put on body armor before going into the confrontation, a prosecutor is likely to argue that you were preparing ahead of time for a fight, and that it's evidence that you were planning ahead of time to get into a gunfight. Which means he may well charge you with murder in the first degree, premeditated murder. Whether the jury buys his argument that you were planning the confrontation is up to them, and the other evidence the prosecution advances.

Perhaps I was unwise to assume that people would read the indictment and that I didn't need to repeat what it contained. So, to be clear: the evidence of pre-planning in the indictment included purchasing rifles ahead of time because "Cops are not trained or equipped for more than one rifle so it tends to make them back off" (direct quote by the guy who ended up sentenced to 100 years for, among several other crimes, shooting an ICE officer in the neck). This suggests that the intent was to use the weapons to shoot at law enforcement officers. Carrying weapons is protected under the Second Amendment, but preparing to commit a violent crime is not.

That's the standard. The only thing that's been shifting is my opinion of whether you're actually looking at the evidence vs. forming an opinion ahead of time and not letting yourself be budged by evidence. Which is why I will no longer engage with you after this comment.


That iMessage rendering looks correct, just typeset differently.

Nice, I have Claude, now I just need an MRI (which in some countries is sort of a hard requirement still unfortunately, long queues)

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

Search: