adamhp

Treat adults as adults? You act like the user base for this are a handful of completely benevolent people. It's getting over a billion monthly visits. It is naive to think that OpenAI should take no steps towards making it safe.

I'm not trying to shut up critics, just the morons calling ChatGPT "woke".

One of the most widely used novel AI technology companies has a vested interest in public safety, if not only for the sheer business reasons of branding. People are complaining about the steps that Open AI is taking towards alignment. Sam Altman has spoken at length about how difficult that task truly is, and is obviously aware that "alignment" isn't objective or even at all similar across different cultures.

What should be painfully obvious to all the smart people on Hacker News is that this technology has very high potential to cause material harm to many human beings. I'm not arguing that we shouldn't continue to develop it. I'm arguing that these people complaining about Open AI's attempts at making it safer--i.e. "SF liberal bros are makin it woke"--are just naive, don't actually care, and just have shitty politics.

It's the same people that say "keep politics out of X", but their threshold for something being political is right around everybody else's threshold for "basic empathy".

I think it helps to come to peace with the fact that the web is extremely diverse, then pick a tool and learn it. In reality, table stakes is HTML, CSS, and at least a general understanding of JavaScript (not a deep dive). Then I'd pick a framework that makes these things a little more intuitive. In my opinion, that's NextJS, React, and Tailwind -- almost solely because there is a pretty large mindshare around them so there are a lot of tools and libraries that support you. Once you get the hang of these things, frontend development becomes pretty easy.

It's not about honesty vs. political correctness, it is about safety. There's real concern that the model can cause harm to humans, in a variety of ways, which is and should be unethical. If we have to argue about that in 2023, that's concerning.

I really really wish people would keep reading 1984.

Sounds like you've done this? Any tips?

The contrast ratio on the body text is 1.45. I genuinely can't read any amount of this. Try this: https://coolors.co/contrast-checker/

Man's Search for Meaning by Viktor Frankl

> Do you actually believe that a technology that has access to vast amounts of human-generated data and can imitate our language well enough to fool us a lot of the time will suddenly scale into or somehow spark itself into an actual intelligence with its own goals and ideas and a drive to survive and reproduce? Or that something that has no body or senses with which to experience the world can evolve itself?

Not in its current state. Have you looked at things like LangChain at all? I think people are overlooking the integration piece of this pie. People are already using it to write code which can recompile. People are already integrating plugins and training it to call functions instead of respond with human language. We're really only a few iterations away from it being able to use what it has learned with human language to be able to emulate "decision making" and that may or may not be reasonable, depending on how it's built or trained.

Despite what people keep saying about "it's just an LLM!", it should be clear to people that language is a human abstraction for meaning, and clearly GPT-4+ can produce pretty meaningful results in the language department. If you give it camera feeds, and access to APIs, and knowledge of burpsuite and kali linux, and on and on, things do start to look a little scary.

Edit: and by give it, I don't mean ChatGPT, I mean whatever the next nth iteration is. I think we're safe for now. :)

The interesting thing about AI is if sending a request to a prompt is your backend, then you can change your backend in a matter of minutes by doing some simple prompt engineering. I think as the consistency improves, it's going to become really interesting. There's actually untapped value there in how it negates the risk of change by making it incredibly easy to update once it changes. This is obviously ignoring any API changes, etc.

Like I recently wrote a little AI wine pairings tool for a fun weekend project (https://oeno.chat) and it was originally GPT3.5, but I've tried similar prompts with GPT4 and it's just a better (although a little slower) version of the same information. I didn't really have to update anything, because I just ask it for specific JSON keys to be returned every time. Every now and then there are some artifacts where parts of the response bleed over, but it's just a toy project.

I'm really excited about leveraging existing tech by gluing it together with AI. Something like: first, get travel recommendations and extract all place entities (give me json); second, take all those and query some external geolocation api; third, build a pretty ui to map those places; last, integrate with Duffel or other booking API to actually book travel.

This is just a random example but I really see the initial leverage being in creativity of gluing together APIs and systems using LLM products to glue things together.

I think there's going to end up being a premium on front-end development until the AI can do better.

> AI is going to allow companies do do a lot more work with far fewer people.

Or just 2-10x the amount of work with the same people? There will be some economic equilibrium, but I don't imagine companies seeing competitors pumping out 10x the amount of work going "Oh we should fire 9/10ths of our staff to accomplish the same amount of work". We're entering unknown territory though, so I'm not sure existing economic models really are going to be able to predict the outcomes and impacts of this enormous technological shift.

I think 2 is correct. I don't think any job falls into category 3, in the long-term at least.

I think all of us should be (a) seeking to add value to the world, (b) seeking to raise the quality of life for all humans. I don't think AI changes that. Once we get over the initial hurdles and hiccups, it will hopefully make it a lot easier. (Some will probably call this idea naive).

I would just continue to learn as much as you can about AI, and not just LLMs.

My friend and I recently built a little weekend project that could serve as some inspiration: https://oeno.chat.

I think one thing to think about is really anything around language. Law, content, customer interaction (email, marketing, etc.). What I really see as the big opportunity here is using it as "glue" to integrate separate APIs and then add UI polish. Doing the thought work of "Oh, if I combine this API with this prompt and then send the result to this other API and use OpenAI to reframe that result and then put a UI on it..." is what is going to make your solution stand out from every other "AI Blog Writer" or "AI Email Spammer" or whatever else.

I'm talking about the design aesthetic that is baked in. The actual results, not the code.

Login screens in the apps I build. I love animating the response. If there's a success, fade out, fade in a welcome screen, maybe a fun spinner... It's the door to your app, so it feels like it's time well spent. It gives the feeling of sophistication if done tastefully and well.

I think a lot of people overlook #1, and it is important. Writing Tailwind feels fast, intuitive, and looks good by default.

It is quite fast to make changes and see immediate visual results. There's basically one less step to making css changes. What you call "lazy" in your OP, the rest of us comfortably call "useful".

If I need to add some padding, I slap on some p-4, my browser refreshes, and I see the result.

I don't really get the spiritual debate. It's a tool, use it or don't.

Correct. The latter point is important, especially in heavily regulated industries, or the public sector. You can't just tell the public you decided to do X because an AI told you so.

No, I've switched most of my knowledge keeping to Obsidian (including plain old copying and pasting links into a list). I sacrifice some fidelity to gain centralization, and it's all version controlled on GitHub, so I can edit it from any browser.

> When we crash you won't need to ask if we're crashing. It will be obvious.

This 100%.

> solving problems involves less lobbying

I think we can all agree this pretty much sums it up. Basically every major problem we face in America is the result of lobbying and gerrymandering.

In many cases, you may want to optimize one of the inputs to achieve the desired outcome. If you think of classic linear regression, you get a rough approximation of how each input affects the output. So, for example, if you built a linear regression model on house pricing, you could be relatively confident that an X increase in number of bedrooms, or proximity to public transport, or some other variable, has a direct affect on your output Y (home price). And in the case of linear regression, you get weights and power for each coefficient, so as a builder you can say "I'm going to build houses closer to the metro to fetch a higher price". Greatly simplified, but I think this is is what is missing from AI. If we're having AI classifying things, it's only one piece of the puzzle to actually do the work of classifying (and that's useful to a lot of people). But the other half of the equation is to explain why things are classified the way they are.

You can kind of prove it is possible, can't you? I mean, we have ourselves, which we're sort of claiming is the ground truth comparison to "intelligence". You can then see that the average human actually have limited intelligence, when you look at say, savants or hyper-intelligent people. Then it must be that some physical structure of people's bodies enables this higher degree of intelligence, and removes the "limit" so-to-speak. The average brain has 86 billion neurons, which we know are mostly responsible for piecing together consciousness.

We also have extensive studies on all the ways we are actually really bad at processing input (a by-product of our primate ancestral heritage). There are entire textbooks on all of the different biases we have built-in. And there are clear and obvious limits to our perception, as well (I'm thinking of the five senses here).

Imagine you're neither constrained on the input side or the processing side of this equation. It becomes kind of a mathematical inevitability that we will be able to create artificial intelligence. When anything can be tokenized and act as an "input", and we can run that through something that can process it in the same way that our brains can, only scaled up 10-fold (or more)...

If there is one thing we're good at, it is thinking that we are the center of the universe. I think that is blinding people to the possibility of AI. We can't fathom it, for lots of good and bad monkey reasons.

Try mentioning Neuromancer and hear about how many characters have eyes like "mirrored glass".

The best way is to be involved in communities focused on this sort of thing. YC and Indie Hackers are two. There is also Indie Worldwide. Aside from that, networking and staying active on Twitter seem to be fairly effective.

I met my co-founder (non-technical) by posting on Indie Hackers that I was looking for a co-founder.

Cooking, reading, playing guitar and making music, video games. Hiking. Walking with an audiobook.

Guitar is the one I wish I was good at. I really enjoy more "complicated" music (jazz and fusion) and all of the players I admire are seriously top-tier. But they've put in the work and I only pick up my guitar a few times a week. Perfectionism is tough!

> the bug [..] only presented when many aircraft at the same time were using the system.

This implies that the calculations are done externally to the plane? That seems... interesting.

It doesn't "know" that the capital of Oregon is Salem. To take an extreme example, if everyone on the internet made up a lie that the capital of Oregon is another city, and we trained a model on that, it would respond with that information. The words "the capital of Oregon is Salem" do not imply that the LLM actually knows that information. It's just that Salem statistically most frequently appears as the capital of Oregon in written language.

Thanks for sharing, I hadn't found a nice semantic nugget to capture these thoughts. This is pretty close! And I've heard of the stories described in the "color terminology" section before.