nsagent

Interesting take on the piracy aspect:

  We will have a trial on the pirated copies used to create Anthropic’s central library and the resulting damages, actual or statutory (including for willfulness). That Anthropic later bought a copy of a book it earlier stole off the internet will not absolve it of liability for the theft but it may affect the extent of statutory damages. Nothing is foreclosed as to any other copies flowing from library copies for uses other than for training LLMs.

Having worked with game designers writing code for their missions/levels in a scripting language, I'd say this has been the case for quite a long while.

They start with the code from another level, then modify it until it seems to do what they want. During the alpha testing phase, we'd have a programmer read through the code and remove all the useless cruft and fix any associated bugs.

In some sense that's what vibe coding with an AI is like if you don't know how to code. You have the AI make some initial set of code that you can't evaluate for correctness, then slowly modify it until it seems to behave generally like you want. You might even learn to recognize a few things in the code over time, at which point you can directly change some variables or structures in the code directly.

Weird. I've lifted on and off for 25 years. For most of that time I did the stereotypical 3x8-12 and saw much slower progression. During the past couple years I switched to a 5x5 plan and saw massive gains in strength, even while I was cutting weight via a caloric deficit (was eating 1500-1600 calories a day, but had lots of protein and adequate carbs).

For reference, I went from a dumbbell bench press of 45lb to 75lb in 4.5 months (5x5). Previously my progress was much slower.

I'll caveat that I've obviously not closely controlled for all factors and I'm an n of 1. Additionally my interest is in having a great strength to weight ratio, rather than being a body builder. I'm a climber and that's an important consideration.

Just ran into the same issue. The start of the flow certainly looked interesting. Too bas I couldn't get further along. I'll try to look at it again later.

Oh wow. Good to know! Thanks for the heads up.

I used to use Signal exclusively rather than Whatsapp, but I've had lots of issues sending media. This has not been a problem with Whatsapp, so I've recently begun to use Whatsapp more. There are also issues with message history that I've encountered on Signal that don't exist on Whatsapp.

If Signal could address these concerns I'd be happy to move away from Whatsapp.

With this news I'll likely need to reassess my use of Whatsapp again.

I have Signal on my phone and laptop. For some reason my laptop desynced from the phone, so my chat history now has a missing block of message history (that exists on the phone). I did nothing obvious to cause that desync. My guess is that my phone updated the Signal app, and I didn't update it on the laptop in lockstep. That's not a great UX, especially since there is no notification that this might happen.

My mom was one of those lucky few doctors that was able to redo internship and residency after moving to the US. When she first arrived she cleaned houses and worked as a waitress while awaiting entrance into a program. This was the early 80s. Of her friends, she's one of the few who was able to go through the gauntlet and become a doctor in the US. Most became other types of medical workers, including my father who became a respiratory therapist.

Contrast that to my cousin who moved to Australia and was quickly accepted as a practicing physician.

Sure, part of the issue is bad hardware support, but most recently I had a Dell laptop for work with Ubuntu on it. It was officially supported, but refused to go to sleep properly, so each morning it would be nearly drained even if it was fully charged the night before.

I had other issues that were not hardware related though. The desktop environment was missing some basic features for things like mouse settings that I had to install community extensions for, which were buggy.

I also had issues with printers, HDMI output, keyboard settings, and more. The list goes on and on. Each was something I spent time on that I haven't had to spend time on with MacOS (it's been a decade and a half since I've used Windows, but I remember it having fewer issues).

BTW, I also dread OS updates on Linux, and that includes server-side. Definitely another pain point that feels more severe than on MacOS.

Anyway, I'm glad that Linux works for some people's usecases, but it feels like it's been in this limbo of quasi-usable for quite a while from my perspective.

It's clear that you're either one of the authors or a friend of theirs. You created this account 8 months ago to comment on another paper [1] that was released by the same authors.

[1]: https://news.ycombinator.com/item?id=41776324

As someone who grew up running Microsoft OSes, starting with DOS, then Windows and who has used a Mac laptop since the Windows Vista days, my perspective on the usability of Linux Desktop is unrelated to it simply being "different from what I'm used to."

Transitioning from Windows to Mac was much more of an adjustment than Linux Desktop. It's just that Linux has too many rough edges. While it's possible I've simply been unlucky, everytime I've tried Linux it's been small niggling issue after small niggling issue that I have to work around and it feels like a death of a thousand paper cuts. (BTW I first tried Linux desktop back in the late 90s and most recently used it as my main work laptop for 9 months this past year.)

Note, I'm more than happy to use Linux as a server. I run Linux servers at home and have for decades. But the desktop environments I've tried have all been irksome.

Note that I'm not mentioning particular distros or desktop environments because I've tried various over the years.

This is a recent trend and one I wholeheartedly agree with. See these position papers (including one from David Silver from Deepmind and an interview where he discusses it):

https://ojs.aaai.org/index.php/AAAI-SS/article/download/2748...

https://arxiv.org/abs/2502.19402

https://news.ycombinator.com/item?id=43740858

https://youtu.be/zzXyPGEtseI

> Meanwhile, software developers spot code fragments seemingly lifted from public repositories on Github and lose their shit. What about the licensing? If you’re a lawyer, I defer. But if you’re a software developer playing this card? Cut me a little slack as I ask you to shove this concern up your ass. No profession has demonstrated more contempt for intellectual property.

I was willing to read until I got to this part. He can shove this drivel up his own ass for all I care if that's how he thinks. What a waste of time.

I remember sitting with the faculty in charge of offering online courses when I visited as an alum back in 2014. They seemed to look at it as a cash cow in their presentation. They were eager to be at the forefront of online CS degrees at the time.

It really depends on the software. I have the top-of-the-line M4 Max laptop with 128GB of memory. I recently switched from Zotero [1] to using papis [2] at the command line.

Zotero would take 30 seconds to a minute to start up. papis has no startup time as it's a cli app and searching is nearly instantaneous.

There is no reason for Zotero to be so slow. In fact, before switching I had to cut down on the number of papers it was managing because at one point it stopped loading altogether.

It's great you haven't run into poorly optimized software, but but not everyone is so lucky.

[1]: https://www.zotero.org/ [2]: https://github.com/papis/papis

Interesting. At game companies I worked at we generally used version control solutions that easily allowed storing code and assets together, such as Perforce and Alienbrain.

Good question. I started having digestive issues a year and a half ago. Went to various doctors and according to their lab tests was told everything was fine, that maybe this was my "new normal." Recently had an infection that necessitated broad spectrum antibiotics and magically my digestion issues have cleared up...

If not for getting sick recently I'd probably still be dealing with whatever gut infection I had that the doctors essentially ignored.

Congrats!

One little nitpick is that while I prefer hardmode for puzzles like this, on mobile (where I prefer to play such games), it's WAY too easy to misclick on the keyboard. I had "8 excess keystrokes," but those were all due to mistyping then deleting a single mistyped character (e.g. "ee" -> "er").

Not sure how to keep the essence of the skill-based approach, without the frustration induced by typing on a mobile keyboard.

Anyway, great job and congrats again!

Thanks for clarifying! Certainly seems like dissembling to me. Wish news organizations actually did the due diligence and reported these kinds of details.

I'm completely unfamiliar with what systems exist, but here's what the article states:

> the White House had authorized the use of Signal, largely because there is no alternative platform to text in real time across different agencies, two people familiar with the matter said.

> Previous administrations, including the Biden White House, did not develop an alternative platform to Signal, one of the people said.

Are you saying these sources are dissembling? Wouldn't surprise me at this point, but just making sure I understand what you're saying.

> According to the White House, the number was erroneously saved during a “contact suggestion update” by Waltz’s iPhone, which one person described as the function where an iPhone algorithm adds a previously unknown number to an existing contact that it detects may be related.

It's interesting that this was the cause. I'm sure we all have our own stories of how UI/UX niggles (regardless of platform or app) have led to unintended behavior.

While I understand automatic suggestions can be helpful at times, when the UX doesn't clearly identify the cues that lead to the suggestion, with a way for a human to confirm it, this type of error is a likely result.

Nope. I did it purely due to long-term intellectual curiosity.

I first pursued "AI" in undergrad during the last AI winter. For example, the only professor who taught neural networks at Purdue was in the EE dept, not CS, and was retiring the semester I was first qualified to study it. There weren't enough seats in the class, and since it was graduate level, I wasn't allowed to take it as an undergrad.

I really tried every avenue I could think of at the time to pursue AI — taking part in Robocup, taking classical AI (also from the EE dept), etc. None of what I was exposed to seemed like it was pushing the the intellectual boundaries, so I instead got into video game AI as a way to pursue AI (a number of famous ML researchers like Demis Hassabis got their start in video games).

When I started my PhD a very tiny group of researchers were looking at text generation, let alone for creative text. The idea was very niche.

Note, I only pursued a PhD after I got an interview at OpenAI in 2017 that made me realize a PhD was likely necessary to pursue research.

I'll bite.

I received my PhD in Computer Science focused on NLP and creative text generation last year and I think the hype around LLMs is ridiculous (academics are no better than industry on chasing hype). They're trained to predict the next token given a context, and that's exactly what they're good at.

How old do you think I am?

> Banning phones won't help, phones are just a surface level symptom of the fact that humanities courses at minimum have become completely fake and professors don't care enough to stop it.

It's not just humanities courses and it also affects "elite" universities.

For the graduate NLP course my advisor taught at UMass Amherst, despite allowing ChatGPT for a take home test (this was a couple years ago), 60% of the students broke into two separate collusion rings (one Chinese group and another South Asian) and copied off of each other. They got caught when the answers were wrong, but in a different way than ChatGPT. Despite the seriousness of the rampant cheating, students were not failed out of the course, mainly because it reflects badly on the University if they fail. My advisor had to go through a lot of unnecessary bureaucracy in the process.

Jump forward to Cornell University where I'm currently a postdoc and grade inflation is real. They had to get rid of reporting course-wide median grade beside a student's grade on their transcript [1] to help combat it.

That hasn't prevented the pressure to pass students with high marks despite abysmal performance. I supervised an undergrad's research one semester as an independent study course. That student did very little work and despite multiple promptings over several weeks, would fail to provide their code for me to help debug and provide code review. I ended up giving them a B+, which is somehow considered "failing". The student even reached out after grades were assigned to beg me to reconsider. None of the students I've worked with so far have had the skills I'm pretty sure I mastered by that time (this includes work with undergrads, master's students, and PhD students). I'm continually shocked by the caliber of students here compared to what I assumed before joining.

I trust professors who've been teaching for decades when they say something has qualitatively changed.

[1]: https://registrar.cornell.edu/grades-transcripts/median-grad...

That's pretty spot on with my experience as well, though I'm a decade older.

> several key supporting points been mostly disproven scientifically (early humans often hunted by running to exhaustion, for example)

This article from 2024 [1] that discusses newly published research seems to refute your claim.

[1]: https://www.smithsonianmag.com/smart-news/long-distance-runn...

It seems odd that as a society it's been deemed a-okay to pathologize a cluster of traits shared by 30% of the population.

I most likely would be diagnosed with ADHD if I sought a formal diagnosis (my ex, who I was with for over a decade is a clinical psychologist and said I most likely have ADHD; my brother is diagnosed with it and my mom's ADHD traits make the both of us look tame). The thing is, I just don't buy into the rampant pathologizing that's so prevalent these days.

It's really a societal problem, and is being treated as a human failing. I contend that it's society that's expecting humans to behave in a way we haven't needed for most of our evolution.

These include: having exacting time standards disconnected from seasons and day/night cycles, having to juggle many more cognitively intense tasks that we need to quickly switch amongst, being exposed to media and technology specifically designed to capture your attention for as long as possible, etc.

I'm also an AI researcher and I'm not hopeful that the hype will die down anytime soon. People have been touting insignificant results through shoddy science for a long time now. They've noticed it works well enough because current ML is still pretty much alchemy (Ali Rahimi's 2017 NIPS Test of Time [1] talk still resonates today) so people rarely spend the effort to effectively refute the bogus claims.

As a result, I've opted out of the system and I'm working on trying ambitious ideas I have to try to upend the current paradigm of training on big data, which is truly insane (by age 4 the most erudite of children have probably only been exposed to 45 million words [2] and yet exhibit vastly more understanding and fluency than any language model trained on a similar amount of data).

[1]: https://www.youtube.com/watch?v=Qi1Yry33TQE

[2]: https://www.aft.org/sites/default/files/media/2014/TheEarlyC...

Thanks! It's hard to describe, but ultimately I realized that neither academia nor industry is willing to invest in moonshot type ideas. Moreover, academics have to worry about funding and tenure (in addition to teaching and service). Funding is easiest when you propose research topics that people believe stand a chance of success. For wacky ideas it's difficult to make that case.

Anyway, I have enough money to fund myself for at least a couple years, so my goal is to make the best of that time and see if I can upend the dominant paradigm in NLP (humans don't need as much data to learn — I'm going to pursue ideas that allow computers to be as data efficient).