andrewedstrom

So curious who the anonymous example is

These are not solopreneurs. They’re entrepreneurs.

You may be interested in a recent AI safety paper by Redwood Research.

In it, they have GPT-4 generate solutions to coding problems, but instruct it to insert backdoors into the solutions some fraction of the time. Then, they explore different ways to use a weaker model (GPT-3.5) to detect these backdoors. Pretty interesting.

[1] https://arxiv.org/abs/2312.06942

Standard Ebooks is fantastic! In fact, I love what they're doing so much that I actually built a little SaaS product on top of their ebook collection.

The site is called Modern Serial, and it lets you read books from Standard Ebooks in 10 minutes a day as Substack-style email newsletters.

https://modernserial.com/

Slight correction: Portuguese Man-o-wars are actually considered multicellular organisms.

They’re weird organisms composed of multiple distinct animals that cannot survive independently and that originate from the same fertilized egg. But they are unambiguously multicellular.

Rectangle has worked well for me in the past, but recently, it's been losing a lot of windows. I often have to restart Brave to get Rectangle to be able to control it again.

Not necessarily, no.

Open source models are already being used for all kinds of nefarious purposes. Any safety controls on a model are easily stripped off once its weights are public.

Usually I love open source software. Most of my career has been spent writing open source code. But this is powerful and dangerous technology. I don’t believe that nuclear weapons should be open source and available to all either.

You contradict yourself

I can't help but think of the similarities between Frankenstein's Monster and AI. Both are human-created intelligences that learn to understand us by observing our behavior and writings.

Just linking to the education tab of her profile is misleading.

Many people in AI safety are young. She has more professional experience than many leaders in the field.

https://www.linkedin.com/in/helen-toner-4162439a/

Execute Program is one of the best tools for learning a programming language out there. I’ve worked through several of your courses on TypeScript and the Modern JavaScript, SQL, and Regex courses as well. Can’t wait for this one.

Thanks for everything you do!

There are official docs, they just aren’t very good.

You could get the exact same alerting benefit with 1/10 the Bitcoin, or less.

There is no tax penalty for moving bitcoin. You should definitely move most of this elsewhere.

Alternative explanation: ketamine isn’t better than placebo for treating depression when given to someone who’s under anesthesia. You have to be awake.

Similarly, I bet that if you gave psilocybin to someone under anesthesia, they wouldn’t get the same benefits as someone who took it while awake.

Not pictured: the cost of switching to DynamoDB when this service inevitably gets shut down in two years.

It makes sense the CEO would either step down or be forcibly removed by the board.

Unity's mishandling of the Runtime Fee policy announcement has caused permanent damage to their reputation. It was a perfect case study in how to undo decades of trust-building in one day.

I follow a lot of game developers online. Every single one that uses Unity today is planning to switch engines for their future games.

Fantastic, clear, and thoughtful essay. This should be added to every list of recommended readings on the subject of AI risk.

I would never work for this company.

I haven’t had any issues with that. However I don’t send any marketing emails, and since people are paying me to send them these emails I think the chance of getting flagged as spam is particularly low.

Haha well, it's no secret that the books are free.

Instapaper users pay to read free articles from the internet in a way that fits into their lives better. Same concept here.

Thanks, glad you like the site! Yes that is correct—all of the books I sell are in the public domain.

I use Sendgrid to send emails, which has worked great so far although it took a while to get my account fully set up.

Re: customers, so far I've had more than 10 and less than 1000. One encouraging sign is that many people are coming back after finishing their first book and buying another.

Serial Reader is great! I came across it a year ago while looking for a service that would let me read classics as email newsletters.

I used it for a couple of weeks. The app is slick and it has some really cool features (like letting you import your own ebooks) but I kept forgetting to open the app and then falling behind. Also the daily reading length of 20 minutes per day felt too long for me, personally.

Ultimately I got it in my head that I'd be more likely to actually read issues if they showed up somewhere that I already had a habit of checking daily (my email inbox), and if the issues were shorter (~10 minutes each) with more fine-grained controls about when they were delivered. So that's what I built!

I actually emailed Michael (the person behind Serial Reader) back in June, hoping to get his blessing to use the word "Serial" in my product's name. Never heard back and eventually I decided to move forward with the name.

Honestly, a pretty reasonable solution to that problem. It's cool that we have the technology to make that work.

I'm sure the top comment here will be something like "this is invalid because no way can you assign a numerical value to code quality! wtf?!"

I'm withholding my own judgement on that.

For anyone curious, the authors are coming up with a code quality score using an open-source tool called SoftWipe[0]. From the paper:

SoftWipe is an open source tool and benchmark to assess, rate, and review scientific software written in C or C++ with respect to coding standard adherence. The coding standard adherence is assessed using a set of static and dynamic code analysers such as Lizard (https://github.com/terryyin/lizard) or the Clang address sanitiser (https: //clang.llvm.org/). It returns a score between 0 (low adherence) and 10 (good adherence). In order to simplify our experimental setup, we excluded the compilation warnings, which require a difficult to automate compilation of the assessed software, from the analysis using the --exclude-compilation option.

[0] https://github.com/adrianzap/softwipe

I remain confused about how Zig makes it to the front page of Hacker News so often. As far as I know, no one is using it in production after 7 years of development.

Are people just really excited about this project? Are the people behind it just really good at marketing?

Honest question: why do we all keep talking about Zig?

https://archive.is/Y1IwY

Yeah…this one makes a lot of sense. I don’t know about y’all but my visits to Stack Overflow have probably gone to about 1/10 of what they were before ChatGPT.

Not according to the article.

“The study estimated that 15% of schizophrenia cases in men could be averted if they avoided cannabis use disorder”

Perfectly stated. His argument that morality follows from intelligence is totally wrong.

I think you’re right: he’s biased because of his employment at OpenAI.

“It is difficult to get a man to understand something, when his salary depends on his not understanding it.” - Upton Sinclair

I’ve had both many times, and with Soylent I experienced the exact same problem you’ve described. I tried to go “full Soylent” in college and I could drink 3000 calories of it and still feel hungry.

A couple of years ago I tried Huel Black and loved it. It’s way more satiating, healthier, and I actually enjoyed the taste. Unlike Soylent, it actually feels like food. I look forward to having it for breakfast each day.

My favorite flavors are Coffee Carmel and Chocolate. I highly recommend Huel Black over the regular Huel. Steer clear of the Cinnamon Roll flavor.