cloudking

1080Ti from 2017 still handles modern AAA games really well, was able to play and finish Cyberpunk 2077 with no issues. Really a remarkable feat of engineering for it's time.

I'm just getting started, thanks this was super helpful!

Not related to local LLMs, but JOI from BR2049 is essentially what Replika is striving for: https://replika.com/

Infact during the onboarding process they ask the user to choose which AI companion movie they related to the most: Her, BR2049 or Ex-Machina. The experience is then tailored to align closer to the movie chosen.

It's quite a terrible app from a product design perspective: filled with dark patterns (like sending the user blurred images to "unlock*) and upsells, but it's become successful amongst the masses that have adopted it, which I find fascinating. 30m+ users https://en.wikipedia.org/wiki/Replika#:~:text=Replika%20beca....

Download the CDN libraries, point to local

Any tips for launching on Itch?

Hackers, just let us try your product immediately

> Is this like a law of the universe I'm not aware of, that you must be able to create a profitable startup that generates tattoo ideas?

That's not what I said. Your question/argument was why build a wrapper when someone can go to ChatGPT and generate a tattoo. You can make that argument for any startup that wraps AI image models. If everyone followed that argument, there would be no startups in this space.

> A profitable product will make use of APIs to do something that a user couldn't do almost just as well by just prompting ChatGPT themselves

Exactly, average users won't know how to steer ChatGPT to generate high quality tattoos. In fact, OP is not even using an OpenAI model, they are using Flux which has no direct consumer interface (from the creators of Flux) and much higher quality image generation.

I wish OP success on his project.

By that logic, no one should build a startup that generates tattoos then?

Speak to average, non-technical users. You'd be surprised how many people have a very vague idea what ChatGPT is capable of. They aren't using it everyday like you and I. Relating this back to original comment, expecting them to know about effective prompting techniques, Stable Diffusion etc is unrealistic.

One of the reasons OpenAI offers APIs, is so you can build startups on top of their tech for average non-technical users.

You're not the target user. Average users have no idea what you're talking about.

Yeah, Midjourney tends to create sci-fi/enhanced looking humans. Flux creates photorealistic.

For whole codebase prompting, you'll have a much better time with https://www.cursor.com/

You can use OpenAI, Anthropic, Google etc as the LLM provider. o1 is supported

https://docs.cursor.com/chat/codebase

Nvidia just announced something similar https://www.nvidia.com/en-us/ai/cosmos/

In terms of realism, Flux is leading the pack currently

I can afford them, I just don't need to anymore. My use cases for stock photos are websites, marketing, landing pages etc. The SOTA image models are sufficient for my use cases and my customers don't care. Infact, they are happy with the quality of AI generated stock photos and appreciate the fast turnaround and lower cost.

Anecdotal, but I haven't bought a stock image since Stable Diffusion was released.

Edit: with Flux, you can't even tell the difference: https://blackforestlabs.ai/

Yep, works great for simple projects.

I'll give it a go, I've also heard Windsurf is quite good.

Personally I've been very impressed with Cursor Agent mode, I'm using it almost exclusively. It understands the entire codebase, makes changes across files, generates new files, and interacts with terminal input/output. Using it, I've been able to build, test & deploy fullstack React web apps and three.js games from scratch.

Sounds like OP hasn't tried the AI IDEs mentioned in the article.

For example, Cursor Agent mode does this out of the box. It literally looks for context before applying features, changes, fixes etc. It will even build, test and deploy your code - fixing any issues it finds along the way.

WFH was the best thing to come out of COVID.

Maybe the algorithms simply concluded that you might interested in things your friends are interested in.. that's part of why you're friends?

If any of your friends searched for something online, and you're connected to them in a social network, they might show you the same ad by proximity.

We're at the tipping point where average users cannot tell the difference between AI generated and human content. It's already very evident on Facebook and Threads, just spend a few minutes on either platform, you'll find a post with an AI generated image or video and tons of human engagement supporting it.

In the spirit of Christmas I prompted it to "guess the movie" for this image https://wallpapers.com/wallpapers/funny-home-alone-h1e9xnttc...

Results: https://pastebin.com/Wpk3PmUq

What makes this particularly egregious is the comment defending Claude and bashing Gemini was generated by ChatGPT

https://linear.app/ has a clean design

LLMs are tools, learn how to use them. You'll be more productive once you learn how to properly incorporate LLMs into your workflow, and maybe come up with some ideas to solve software problems with them along the way.

I wouldn't worry about being replaced by an LLM, I'd worry about falling behind and being replaced by a human augmented with an LLM.

IMO "Agents" are a marketing term, they are simply software that use LLMs somewhere in the backend. Often daisy chained into a series of operations that may involve additional LLM calls or calls to other internal/external services.

One we've been using for meeting notes + action items works quite well https://fireflies.ai

Canvas was releaased on October 3, 2024 - is this something different?

When crafting projects from scratch, does your system actually fix it's own errors?

That seems to be the challenge with Cursor Agent in it's current form, it generates a bunch of code that has bugs and requires a lot of iteration.

Contains MSTR and not GOOG...

The answer is yes and no, simultaneously