I wanted to create a thread where those that submitted applications to Converge-2 could share what they pitched and whether they received and interview. I personally feel like all of this selection should have happened in the open to avoid any perceived bias… So I’ll go first:
Our pitch was for my new startup Awarity.ai. Awairity is a private data reasoning engine capable of answering complex questions over large corpuses of data. Our website mainly focuses on initial scenarios attractive to a range customers we’ve been working with. At the core of Awarity, is a reasoning engine named Awareness. Awareness is essentially a compression engine for reasoning. The Gen 0 version of Awareness is able to reliably compress around 500,000,000 tokens of content down to around 8,000 tokens of reasoning. The new Gen 1 version of Awareness I’m working on will increase that compression rate to billions of tokens down to the same 8,000 tokens of reasoning in even faster time using a slew of new techniques I’m developing.
Winning converge would help to further accelerate are work on Awareness but if we don’t win its not the end of the world. We know we have a billion dollar business in the making.
So far we have not been invited for a review.
Please share your project. I’m eager to hear all the great ideas people have submitted and not just the 10 that will be accepted.
Hey, stranger! Good to see things are going well with Awarity.
We’re making progress… But I’m really excited to hear about everyone elses projects.
Oh, I’m tinkering here and there and everywhere. Haha.
I’m really hoping to finish up the AI Roguelike project soon… Livestream from last week…
I hope you get chosen with us! My company could absolutely work with you.
We have created a group chat environment and document tool where multiple agents can be requested to respond or create following Professors commentary or work. We have been testing it at the college i am studying and have a separate agent dedicated to finding keywords for restructuring documents using some odd solutions through Microsoft semantic kernal and some new ways to use basic JSOn functions.
Being able to token down massive articles or documents and data would be HUGE for helping our model better create a document during the conversation/group chat.
I would love to see you guys get called on board. And even if you don’t I wouldn’t mind working together on something in the future!
Did you guys get selected for an interview?
I have my own AutoGPT style Agent Framework called AlphaWave agents but I’m not using it for what we’re building. Traditional agent frameworks won’t work for what I’m doing so I had to go back to first principles. The issue is the first principles for this level of compression don’t exist yet either so I had to invent them (which took me 3 months) but I have it down to general purpose compression algorithm that I’m not quite ready to share the details of yet. I know that sounds like hyperbole but everyone on this site that knows me also knows that when I say I have something working… I have something working.
There’s only 10 spots and thousands of applicants. We shouldn’t take it personally if we don’t get selected. Awesome if you do get selected but it is a bit of a lottery.
I applied with a startup using a robust recommender system. Anyone is proficient here in recommender system?
Are you able/willing to expand on what you mean with compression in this context?
Hi @stevenic, we applied with our system that automates the process of metadata tagging and tokenization. Our focus on content files incl old adobe formats etc. We are helping companies structure content data, to feed AI training sets.
Hi @Niamh , does your content data format includes / compatible with 3d ? Like USD or alembic ?
Not yet! We are looking to increase the platforms capabilities to include CAD files and other 3d and data rich file types. We need the money to invest in development - hence the reason for applying to Converge 2!
That’s awesome ! Have you got interview schedule yet from Converge jury board ?
I think 3D files support would open lots of interesting applications and use. Look for USD, Alembic, FBX, and other files , with animation data.
Not yet - I doubt we will get one considering the length of time since applying!
Thanks so much for the suggestion @raiyan – we will definitely add them to the list of file types.
Is there anyone here that is already being contacted for the interview? I cant find any dates about that
Did anyone received their confirmation? I’ve pitched in for Diagnostics use case.