"AI for good, for real.".
And we mean it!
The problem with AI today
Generative AI is extractive. It is trained on work taken without consent, it wastes enormous amounts of energy to produce and utilize, the internals are largely inexplicable and opaque, and it is owned by a handful of companies who sell it largely without a conscience.
We believe you should have better options, those that consider the harms wrought by the industry at large, and those that do the best possible job to mitigate them.
We care about the environment.
Do you think that the world's biggest tech companies lack the resources to solve the environmental problems they create? We don't. It is a choice. We build efficient AI. Our vision is to create models that can run on commodity hardware, powerfully, and cleanly, without need for large use data centers or specialty hardware.
We also point our models at problems worth solving: improving climate change research, and helping scientists do tough, intractable jobs more efficiently. All of our models utilize the Hippocratic license for a reason. We will never release a model to "replace" a human or to generate creative work.
Our first goal on our roadmap is to build out this vision - what would truly minimal, efficient, and general AI look like? Could we use it as a force for good? To help advocate for creatives? To center the humanity of creative output again? We know we can.
We care about people.
Consent comes first. We will find a solution to pay creatives fairly for their inputs. We aim to construct clean datasets with safe content, with consent, preventing liability for the industry at large. We think it should be the norm that provenance is tracked. The people whose work makes our models possible should receive the benefits of the added value.
Our second goal on our roadmap is to figure out how we can use our models to provide restorative justice. Can we establish a data clearing house? Can we use our ultra efficient AI systems to help label data lineage and provenance - for an industry that cannot generally certify, beyond doubts, that it is using ethically procured data sources? Can we then use that to pay creatives fairly? One thing is for sure - we will take on that challenge even if nobody else joins us.
We will not repeat the same tired patterns
We cannot get to a better outcome by copying the playbook that created the problem. At Lowdown, we are fully and radically transparent, equal, and democratic. DEI is the core of how the organization is built, not an afterthought. We want to the right things.
We will soon be structured as a democratic public benefit corporation so the incentives stay aligned with doing good, rather than with extraction. We will always put our money where our heart is.
Fair trade for AI
Think of us as the fair trade, B Corp hopeful, AI company with something to prove. Maybe we cannot clean the slop from the world. What we can say - we promise to do our best to be excellent, to shine a light on the darkness, and never back down from the truth. Better is possible. This is the way.