





we founded a network out of necessity. there are unknown unknowns in any entrepreneurial career, but that shouldn't stop you from achieving success. so we gathered the resources, the knowledge, and the capital in one place.
we estimate the true technology or manufacturing readiness of the company, identify its constraints, and work in a partnership capacity towards solutions.
we build up companies to a stage ready to reach a capital raise. we take no equity and no percentage of any introduction, because to us it stifles the forward progress of a company.
why do we do this? we've been where you are now. the most invaluable thing is a network of founders that reach out and assist.
the primary barrier stopping anybody from the next breakthrough is a lack of assistance. it takes a community to raise a company. we're here to help.
| project | type | details | released |
|---|---|---|---|
| phone-agent typescript agent stack for phone-centric workflows with docker-friendly deploy paths. | agent | typescript agent stack for phone-centric workflows with docker-friendly deploy paths. | 04/03/26 |
| dandelion cuda / python / common lisp workshop for neuro-symbolic prototypes. | research | cuda / python / common lisp workshop for neuro-symbolic prototypes. | 04/02/26 |
| bently coder 7b qlora coding model on qwen2.5-coder-7b-instruct; bigcodebench hard 92%, humaneval 86%. | model | qlora coding model on qwen2.5-coder-7b-instruct; bigcodebench hard 92%, humaneval 86%. | 03/01/26 |
| dither desktop dither dock: floyd-steinberg, ordered, atkinson, and shape dithering. | tool | desktop dither dock: floyd-steinberg, ordered, atkinson, and shape dithering. | 07/08/25 |
a shared nvidia dgx spark, open to the community. it serves open models through an openai-compatible api.
read the planwe measure progress in decades. to make sure our vision for the future succeeds, we accelerate founders in days.