Where 'hard tech' is defined as requiring lots of time, money (or both) and may or may not be possible with high technical risk.<p>I'm just really interested in people working on fixing hard problems and want to know what is out there.
Have you ever worked in a distributed system? If not, try one. If yes but you don't feel it hard, try improving its reliability/availability/scalability/performance/latency/debuggability/observability/consistency one order of magnitude (e.g. one more 9 in addition to 5 9s) at a time with one order of magnitude more distribution (e.g. spreading a global service to 10x more clusters/data centers).
Founding: Genteract (<a href="https://www.genteract.com" rel="nofollow">https://www.genteract.com</a>).<p>Goal: Real personalized medicine, including predicting which patients will respond to a drug, which won't respond well, and which will experience serious side-effects, based on their genetics.<p>We actually have a working system that can do this today: it's a new genetic analysis methodology that finds all the SNPs associated with a Gene-Environment interaction (GxE) through analysis of clinical data, and generates predictions for how other individuals will respond to the same environmental stimulus (food, drug, behavior, etc).<p>There's a long timeline and potentially large expenses involved in getting the right data to do drug predictions on (drug clinical trial data or EMR data), performing prospective clinical trials with the generated predictions, and finally getting FDA approval.<p>So we decided to start by analyzing existing NIH clinical study data (which we have access to by permission of the NIH and the respective study managers), focusing mainly on interactions between foods and nutrients (as Environment variables) and health parameters like BMI, sleep quality, cognitive measures, heart rate, etc. (as phenotype variables).<p>We're gearing up to launch a service that gives people access to these (and future) discoveries through analysis of their genetic data (either 23andMe or Ancestry genotype files, or whole genome sequencing that we'll offer).
Everything related to drone racing. Applications could be widespread -- but it's also just fun to watch/do.<p>Some research related to psilocybin, ibogaine, and MDMA happening, mostly in Canada. Several of those firms have gotten lots of small (<8MM USD) investments and look interesting. High risk, long-term plays.
Building learning systems that can operate on multiple modalities, and are totally interpretable. I think of these requirements as the basis for the next big jump in software usability, (I.e much better intelligent user interfaces)<p>AGI is not what I want to build, someone else can do that. For now I want to increase the capacity for people to build neural networks that researchers don’t dream of doing, as easy as stitching together web APIs. I internalize this under the title “the Infrastructure of Intelligence”<p>Also, eventually I would like to work on a programming language for biology, and contribute to building wetware computers.<p>Another thing that I think the first project would help is longitudinal health tracking and quantifying human biology.