Hacker News Reader

devxpyon LLMs aren't world models

I don’t think you even need such complex tests as chess to see it doesn’t have a world model - just ask it to do any 5+ digit multiplication

devxpyon I'm Peter Roberts, immigration attorney who does work for YC and startups. AMA

Hi Peter! Heard your name from a couple of friends who have gotten their visas from you! Nice to see you here.

What kind of travel restrictions apply for someone on an O-1 visa (potentially looking to file for EB-1A)? Is there a recommended minimum period of stay in the US?

devxpyon Google to invest up to $2B in Anthropic

The main issue is that its very slow and expensive to browse the internet like this. The LLM will only perform well if you have it do chain of thought reasoning, and that has a latency hit because of a longer generation.

devxpyon Google to invest up to $2B in Anthropic

I’ve just the thing for you my friend https://github.com/devxpy/anti-chatgpt

devxpyon Ask HN: Who is hiring? (July 2023)

Gooey.AI | Full Time | REMOTE / ONSITE | Bengaluru | http://gooey.ai/

We are building a GenAI platform that abstracts all the latest LLMs, animation tools, speech reco, etc from OpenAI, Stability, Google, Meta, etc. One of our projects - https://farmer.chat - was demoed at the UN General Assembly in April 2023.

You'll be the 2nd engineer alongside me - https://github.com/devxpy

Anyone with experience building developer platforms, no-code solutions, pretty UIs, AI models, k8s GPU clusters, prompt engineering, reach out!

Here's a list of models we currently feature, they're pretty cool! - https://gooey.ai/explore/

JD - https://www.help.gooey.ai/jobs/senior-software-engineer

Email - hireme@gooey.ai

devxpyon React, but in Python

Don't use pickle. Use JSON for stuff you can't represent in HTML forms.

For database connections, large datasets, etc (stuff you can't send off to the browser) use redis.

This is a trade-off between DX and UX. Holding session state means you need sticky session routing, and restarting servers kills your user's sessions. Plus imposing a websocket on your users is a cardinal sin, makes scaling incredibly hard, makes page load times abysmal, makes disconnects a nightmare for both you and your user.

Its fine if you're creating something like stable diffusion web ui though, which is meant to be a single user app.

devxpyon React, but in Python

The way to solve session management is to just store the state in html forms and send it back to a stateless http endpoint. That way you get a website that behaves the way all websites work, and can scale easily.

devxpyon Reflected Diffusion Models

Is there a way to run this locally?

devxpyon You probably shouldn't use OpenAI's embeddings

https://gooey.ai/doc-search/?example_id=8ls7dpf6

No code needed :)

devxpyon Using GPT3, Supabase and Pinecone to automate a personalized marketing campaign

Creating the index on pinecone takes about a minute. Creating a table on postgres should take a few milliseconds!

devxpyon The GPT Architecture, on a Napkin

Would also love this to include InsturctGPT architecture with its RL reward model!

devxpyon Ask HN: Who is hiring? (December 2022)

Gooey.AI | Full Time | REMOTE / INTERNS / ONSITE | Seattle & Bengaluru | http://gooey.ai/

We're building a Generative AI no-code platform. Think Zapier, but for Software 2.0

Anyone with experience building developer platforms, no-code solutions, pretty UIs, AI models, GPU orchestration, please reach out!

Here's a list of recipes we currently feature, they're pretty cool! - http://app.gooey.ai/

Email - devs@gooey.ai

devxpyon Show HN: Stable Diffusion v2 web interface

Yeah, seems like its been hugged to death a bit. A single GPU is not quite enough.

Completes in around a minute for me right now

devxpyon Show HN: Stable Diffusion v2 web interface

Shameless plug - We don’t require logging in!

https://app.gooey.ai/CompareText2Img/?example_id=1uONp1IBt0Y

devxpyon Stable Diffusion 2.0

Try it out here - https://app.gooey.ai/CompareText2Img/?example_id=1uONp1IBt0Y

devxpyon Stable Diffusion 2.0

Fixed. Thanks for reporting.

Here's the bug that caused this too - https://bugzilla.mozilla.org/show_bug.cgi?id=1689099

devxpyon Stable Diffusion 2.0

Hmm, looks like we didn't test on Firefox!

devxpyon Stable Diffusion 2.0

I've also built a similar service[1] that does this with inpainting instead of textual inversion, so it preserves the face exactly and returns in seconds, not hours.

[1] - https://app.gooey.ai/FaceInpainting/

devxpyon Show HN: Create art from your face using Stable Diffusion

Thanks, this is helpful!

devxpyon Show HN: Create art from your face using Stable Diffusion

It does the masking automatically by doing face recognition.

That's why the mask and cutout are part of the output too!

devxpyon Show HN: Create art from your face using Stable Diffusion

By others do you mean the people whose code/models I'm using?

devxpyon Show HN: Create art from your face using Stable Diffusion

Added.

devxpyon Show HN: Create art from your face using Stable Diffusion

Adding a privacy policy. "not that it would help much" -- Any ideas on what would help much?

devxpyon AWS vs. GCP reliability is wildly different

Oh interesting. Which region and GPU type were you working with? (Asking so I can avoid in future)

devxpyon AWS vs. GCP reliability is wildly different

Is this testing for spot instances?

In my limited experience, persistent (on-demand) GCP instances always boot up much faster than AWS EC2 instances.

devxpyon Ask HN: Who is hiring? (August 2022)

https://dara.network/ | Senior full stack engineer | Full time | Remote (Anywhere) or Onsite (Bangalore, India)

We're on a mission to build communities for changemakers. Our app, Dara is already used by thousands of the best artists, change makers and institutions in the world to connect to each other and their alumni. Dara is inspired by the need for all of us - but especially those trying to make change - to have a safe, private digital space to discover the best work and communicate authentically within our communities.

Tech stack - You’ll be working with Flutter (Dart) as our primary client side language and Python (Django) on the server. We also use HTML/JS/CSS for rendering SEO content, and of course host our own servers for everything.

Perks - Flexible working hours (except 1 daily standup) + office setup budget + health insurance We’re also very small (2 engineers) so you get tons of autonomy.

Get back to me at https://dara.network/dev/ with your coding superpower, and we'll continue the conversation there.

Looking forward to hearing from you :)

devxpyon Vite – Next Generation Front End Tooling

Try flutter, its a breath of fresh air!

devxpyon Ask HN: Who is hiring? (July 2022)

https://dara.network/ | Senior full stack engineer | Full time | Remote (Anywhere) or Onsite (Bangalore, India)

We're on a mission to build communities for changemakers. Our app, Dara is already used by thousands of the best artists, change makers and institutions in the world to connect to each other and their alumni. Dara is inspired by the need for all of us - but especially those trying to make change - to have a safe, private digital space to discover the best work and communicate authentically within our communities.

Tech stack - You’ll be working with Flutter (Dart) as our primary client side language and Python (Django) on the server. We also use HTML/JS/CSS for rendering SEO content, and of course host our own servers for everything.

Perks - Flexible working hours (except 1 daily standup) + office setup budget + health insurance We’re also very small (2 engineers) so you get tons of autonomy.

Get back to me at https://dara.network/dev/ with your coding superpower, and we'll continue the conversation there.

Looking forward to hearing from you :)

devxpyon Blacklist is not a racist word

I was wondering why all of a sudden the default github branch was “main”

devxpyon iOS 16 to support Web push notifications in 2023

Incredibly sad state of affairs :(