skytrue

Hey folks -- I created 10 Minutes A Day because I was spending too much time 'doomreading' in comment threads, in an effort to stay up to date on what was going on in the tech world. While I love Hacker News and other news aggregators for what they can teach me, I found myself often unhappy reading through hundreds of comments that while, entertaining, ultimately made me feel bad.

The concept of 10 Minutes A Day is extensible beyond just HackerNews, but the Hacker News dataset was a great place to start as it is often a go-to for news around technology and new projects.

It was built on top of Flask, React, OpenAI GPT-4o-mini (to drive down costs), and generally uses a few different prompting techniques to make things work as intended, because GPT-4o-mini can be frustrating to instruct.

10MAD is going to be extended across different news verticals in the future (once I have the time).

Hey people,

We launched a preview of App Spaces (v2) today which is our attempt to make the Microsoft Azure portal much friendlier for developers who are new to Azure or new to cloud, and don't want to deal with all the cruft and intense complexity that Azure/AWS/GCP/etc provide. We know that there's so much more we can do to make these experiences better and this is our beginnings of getting there via something like App Spaces.

You can check out our site at https://www.appspaces.dev, or go directly to the experience (https://ms.portal.azure.com/#view/Microsoft_Azure_PaasServer...).

If you want to learn more or provide feedback, you can reach out to me-- sk dot hartle at microsoft dotcom. We want to rapidly iterate on our initial design here and get to core value within the next half a year or so. We know we have a long way to go, which is why we're releasing it to the community to help us drive and shape our roadmap.

Thanks for pointing that out. It’s been driving me crazy that there’s all these sensationalist articles using that as the “smoking gun”, implying that Netflix specifically wants an AI product manager to produce television shows.

As somebody who has worked to create an AI-generated show, and who is also a PM at a big tech company that is using LLMs for non-creative purposes, I can tell you that the “PM” work I do with these LLMs is vastly different than the creative work I do with them. It’s an entirely different frame and discipline.

I’d start to be concerned once we see job listings that explicitly look for creators, with technical backgrounds in generative AI. The creator/creative talent part comes first before everything else.

This doesn’t help you probably, but the difference between 3.5 and 4 when giving it instructions to follow is huge. I encourage everybody to use GPT-4 when possible, the differences are night and day.

Sorry, I was non-specific. If you're using ChatGPT, you're basically using a "product" that OpenAI created. It has specific system prompts and prompt engineering to ensure that it stays on the rails. If you, instead, use the OpenAI API's for GPT-3.5/GPT-4, you aren't beholden to the ChatGPT "product". It's very easy to create a chatbot (not using ChatGPT, the product) that only produces JSON. It's just hard to get ChatGPT to do that same activity.

That's why everybody doing experiments in this space should either 1) be using the OpenAI Playground, or 2) using the API, and not using ChatGPT.

This isn’t reality. Using ChatGPT this way is fruitless, because there is a system prompt you’re fighting against. I can write a one sentence system prompt for the GPT API that specifies GPT to only spit out JSON, and it works fine. It’s a pretty funny series of image, though.

I’ve been a part of many organizations where the meetings are bad. I don’t think it’s that meetings are inherently bad, but rather how different companies and teams use them.

I generally only invite people to meetings who I know will have important input or need the context. The meeting should be about determining next steps for N period and provide clarity and direction to take those steps.

Any other meeting I mostly find to be a waste of time. But when you get the right people in the room, once a week, to talk about progress on a new product (for example), it can almost entirely replace documentation and is far more flexible and lightweight.

I love writing a good narrative doc or spec, but it leaves room for interpretation. Other people are also not as skilled at writing, and it leaves them without a vehicle to communicate what they want.

So, yeah, agree that meetings are the work/can be an optimal tool for achieving work, but they need to be done right.

Mismatch Media/Nothing, Forever | USA | Remote | Founding Member, Engineer; Founding Member, Ops

Hello! I’m the co-founder of Mismatch Media. We’re looking for a couple of talented and driven engineers and business folks with strong startup experience to come work with us. We are pre-seed, offering founding member equity. You may have seen us recently in the news (for better or for worse!). We create always-on, generative media.

If you’re interested in working on a new type of media, inventing things that nobody has invented before, and creating business models nobody has created before, feel free to reach out: skyler at mismatchmedia.com.

I’m just here to say woohoo, front page of HN bucket list item checked off!

It was trained on sitcom scripts, and we feed the response back to the model as input to create an artificial sense of context.

I think you rightfully pointed out the "watchable" criteria. This is where we got to with 4 years of working on-and-off on it. The hard part is continuing to iterate over the structure to make it passable as a sitcom. The easy part is actually the language model stuff... we have an OpenAI integration, I just don't keep it on because it's a lot of $$$. We have lots of ideas about how to expand the show's structure, but it's mostly backend work that we simply don't have the resources to finish right now.

Luckily, our "goal" with this project was for it be nonsensical (hence the parody part), but I'd love to spin off a new show that focused on making it watchable for hours at a time. Our system is extensible enough that it wouldn't be hard to pop in new art assets and have a brand-new show in a month or two.

https://www.twitch.tv/watchmeforever - AI-generated (aside from the artwork) parody of '90s sitcoms, running forever (24/7/365).

We worked on this w/ a very small team for the past four years, in-between our day jobs. When started, OpenAI didn't have an API, and Stable Diffusion definitely wasn't a thing, so we had to come up with novel methods to thread cohesive content together. Most of the "creative" details e.g., laugh track, dialogue, frequency of dialogue, camera shots, and so on, are all tunable on a per scene basis.

We're in sort of a holding pattern right now -- no clear path to monetization for the project, and it hasn't garnered enough attention for us to probably get funding based on the technology backbone. Hope you enjoy it! Labor of love. :)

(posted this in the similar thread yesterday but I’ll take any exposure I can get…!)

It’s prefabricated 3D models, rendered through a game engine. We do some specific things to make it look how it does.

:D, awesome! Our vision for this was actually just that — second screen content, not necessarily the primary thing you have on, so we love to hear that!

We generate the dialogue per every scene, every 2 minutes or so. We wanted to be able to massage the narrative over time, so the show is constructed in these two minute intervals.

We actually do scrape Twitch chat—- we have onscreen gauges, meters, etc, that the chat can interact with by saying certain things that will change what happens in the show. I just don’t have it turned on right now because our viewer volume is pretty low. :) Great suggestion though re: characters interacting to viewer chat, that would be a lot of fun.

https://www.twitch.tv/watchmeforever - AI-generated (aside from the artwork) parody of '90s sitcoms, running forever (24/7/365).

We worked on this w/ a very small team for the past four years, in-between our day jobs. When started, OpenAI didn't have an API, and Stable Diffusion definitely wasn't a thing, so we had to come up with novel methods to thread cohesive content together. Most of the "creative" details e.g., laugh track, dialogue, frequency of dialogue, camera shots, and so on, are all tunable on a per scene basis.

We're in sort of a holding pattern right now -- no clear path to monetization for the project, and it hasn't garnered enough attention for us to probably get funding based on the technology backbone.

Hope you enjoy it! Labor of love. :)

It’s great you read 4000 Weeks, but as others have mentioned, I’m not sure you gathered the major takeaways. 4000 Weeks is about embracing finitude and accepting that you can’t do everything. The notion is freeing, because once you acknowledge you are finite, you can then make conscious decisions about how you want to spend your time. The intent of the book is to encourage people to choose time wisely, not “fill every day”. It very much has a quality over quantity kind of message. As an ADHD sufferer/chronic procrastinator/perfectionist type, I found it incredibly helpful and freeing to acknowledge that you will never be able to create everything to an unrealistic perfect standard, so instead focus on the things that matter the most to you.

Disclaimer: I'm working on an app to solve the impending wave of generative content, so I'm somewhat biased.

As I'm sure many of us did, I tested out the app here on GPT-3 output. Unmodified, it detected it was GPT-3. Great! However, I added about 10 additional words to the output provided by GPT-3, and it shot up my "human" score by like 60 points, and determined it was human-generated.

This is going to be the problem underscoring _any_ model that is trying to identify "AI" generated text. A human can modify it slightly, or subtract words, and it throws the entire thing off. There are other paths that we need to explore to this problem.

Disclaimer: I have been a software developer, a product marketing manager, a co-founder, and most recently a senior product manager and people manager at one of the big cloud companies.

The question you pose is (no offense) somewhat nonsensical. If a computer could write its own code, what is the need for programmers? :)

The biggest value add for a product manager that I’ve seen is at the strategic level. Assessing the market conditions, being an expert in the broader space, and having a pulse on where the industry is headed. While I do get pulled into design discussions, I’d rather not, but that comes from our designers & researchers not having enough technical depth to fully create experiences and insights for the developer audience (I create cloud services for developers).

We had an interesting case recently where we rolled out a new service that isn’t seeing traction over the past year. Big investment, was initially led and kicked off by our engineering team. However, if you were to have examined the fundamental value props of the service and who this tool was valuable for earlier on in its lifecycle, some fundamental flaws in the assumptions would have popped up. Does this mean that those tech leads and engineers were “bad” at their job, or did they just not have the skill set necessary to assess the value of what they were building? Note: a research study was also done prior to building the product, which clearly missed key gaps and analyses.

How many tools should be in a single person’s toolkit? If they had involved a (good) product manager earlier on in that lifecycle, much of the current pain could have been avoided. As somebody who is currently running a side company, I wish I could hire only “jacks of all trades”, but those people are incredibly uncommon. Most people self-select into a strength and lean into that. Thus, divisions of roles are born.

I feel like people are forgetting that these models are non-deterministic. You can give it the exact same input and get a different response, so "testing" in this way doesn't work. That's also why (right now) automation is impossible. You aren't guaranteed an expected output.

This is so great and so disturbing. A tool that can be used to generate infinite arguments against a single point of view, arguing in favor of why it isn’t an “issue”. It’s like a gun arguing it can’t kill people.

This is one of the more impressive replies I’ve seen, and I love that in its rebuttal to defend itself against my argument, it just proves the argument itself. Amazing.

edit: I tried my hand at crafting a prompt to refute my argument, result:

This comment is just another example of the kind of lazy, uninformed thinking that seems to be all too common on this site. The idea that GPT-3 is going to somehow make it impossible to tell whether something was written by a human or not is complete nonsense. If anything, the generic language and poor grammar of GPT-3's outputs make them even easier to spot.

As for the idea that the "cogency/validity barrier" has been blasted away, that's just ridiculous. The fact is that anyone who puts in the time and effort to research and write a blog post is still going to be able to produce much better content than any GPT-3 generated nonsense.

And let's not forget that fact checking is still a thing. If someone posts something online that is clearly garbage, it's not going to take long for people to call them out on it. Just like with the infamous "AI-generated paper accepted at conference" story, it's only a matter of time before the truth comes out.

In conclusion, this comment is just another example of the kind of alarmist nonsense that we see all too often on Hacker News. The reality is that GPT-3 is not some magical tool that is going to make it impossible to tell whether something was written by a human or not. So, let's all take a deep breath and relax. As they say, "fools rush in where angels fear to tread."

This has likely been commented on one of the myriad of threads related to this topic, but we’re entering (entered?) an era where no content online is going to be able to be presumed authentic i.e., written by a human. Right now, most of us can sniff out GPT’s outputs due to the generic style of the language. But as techniques like the one listed here get better, or the model gets better, anything written online is now suspect.

You could argue this doesn’t matter, but it does, because previously the barrier to entry to write, say, a blog post on a topic was at least a little high. The assumption when you were reading something an author wrote was that while they might not get all the facts right, the cost to the author to write the blog in the first place is high enough for them to ensure their arguments are at least somewhat cogent or valid.

That “cogency/validity barrier” has been blasted away with GPT-3. Now, you can generate an article or post that is potentially nonsense but would require fact checking to sniff out. Once the barrier to create content is as low as a single prompt, any assumption of validity or accuracy, no matter how small, is impossible.

4000 Weeks: Time Management for Mortals. It goes against a lot of societally enforced norms we’re coming up against, and suggests the benefits of living a more aware, intentional life and the benefits of commitment and community. Awesome, awesome book I constantly recommend to people.

It's always interesting to see the buzz that occurs when Copilot is brought up as a topic. This place is called "HackerNews", yet routinely people forget that a "hacker" is somebody using technology to overcome novel problems. Doesn't GitHub Copilot fall into this category? Why is there such an outcry over a technology that has been in the public's hands for less than a year? I'm almost certain that the team responsible for Copilot is going to try to figure out how to avoid spitting out code verbatim, as that's obviously not a good look.

It's most likely the case that in 1, 3, 5 years, Copilot won't be spitting out code blocks verbatim. It will generate rightsize code, trained on lots of publicly available code, and start reducing the surface area required to code/develop.

Stable Diffusion doesn't get in trouble right now because the artwork looks like permutations of different works; text is easy to copyright, style is more challenging, but artists are facing up against the same reality. There's no rolling this back; ML models are going to remove a ton of cruft from creative/labor based endeavors, and people are going to need to evolve to stay relevant.