Andreessen Horowitz Leads $11 Million Seed Round in Distributional

<p><strong>SAN FRANCISCO<&sol;strong> &&num;8212&semi; Distributional has raised &dollar;11 million to build an enterprise platform for artificial intelligence &lpar;AI&rpar; testing and evaluation&comma; with the goal of making all forms of AI safe&comma; secure and reliable&period; The Seed round was led by Andreessen Horowitz with participation from Operator Stack&comma; Point72 Ventures&comma; SV Angel&comma; Two Sigma&comma; Willowtree Investments and dozens of AI leaders as angel investors&period;<&sol;p>&NewLine;<p>&OpenCurlyDoubleQuote;I directly experienced this testing problem while applying AI at Yelp&comma; optimizing models for customers at SigOpt and running a hundred-person AI &amp&semi; HPC engineering team at Intel&comma;” said Scott Clark&comma; Co-Founder and CEO of Distributional&period; &OpenCurlyDoubleQuote;I learned that to robustly test AI I needed to evaluate distributions of outcomes and that there is no purpose-built software for this task&period;”<&sol;p>&NewLine;<p>AI is complex&comma; unpredictable and constantly changing&period; Whether due to hallucinations&comma; instability&comma; inaccuracy&comma; or dozens of other potential challenges&comma; it can be hard to identify&comma; understand and address AI risk&period; To meet this challenge&comma; some AI product teams rely on insights gathered during training that rarely translate to model behavior in production&period; Others rely on monitoring to quickly catch errors in production&comma; but this leaves their customers exposed to potential harm or a poor user experience&period; And some teams run bespoke tests on their models prior to production&comma; but these tests are inconsistent&comma; incomplete and insufficient&period;<&sol;p>&NewLine;<p>&&num;8220&semi;Lack of reliability in AI systems is one of the biggest barriers to widespread enterprise adoption&comma;” says Martin Casado&comma; General Partner at Andreessen Horowitz&period; &OpenCurlyDoubleQuote;We are excited for Distributional to address this problem by building a platform for robust and repeatable AI testing&period;&&num;8221&semi;<&sol;p>&NewLine;<p>Distributional is working with more than a dozen design partners to build an active testing platform that makes it easy for AI product teams across finance&comma; technology&comma; energy and manufacturing industries to get a complete view of AI risk&period; The platform will handle all model types&comma; including statistical models&comma; machine learning&comma; deep learning&comma; large language models and other forms of generative AI&period; With Distributional&comma; AI product teams will continuously catch and address issues before production&period;<&sol;p>&NewLine;<p>&&num;8220&semi;A number of AI product managers that I have spoken with have told me models are failing in production with increasing regularity&comma;&&num;8221&semi; says Noah Carr&comma; partner at Point72 Ventures&period; &&num;8220&semi;As a result&comma; I believe generative foundation models are becoming more critical&period; As demand for implementations grows&comma; so does the potential risk that applications leveraging these models will be pulled offline due to issues related to model shift or exposure to misinformation&period; We are excited to back Distributional’s efforts to enable these teams to catch such issues before their customers do&period;&&num;8221&semi;<&sol;p>&NewLine;<p>Distributional was founded by CEO Scott Clark and an 11-person founding team with experience testing complex AI systems at Bloomberg&comma; Google&comma; Meta&comma; Intel&comma; SigOpt&comma; Slack&comma; Stripe&comma; Uber and Yelp&period; Scott previously co-founded the pioneering AI startup SigOpt&comma; which was funded by Andreessen Horowitz in 2016 and acquired by Intel in 2020&period;<&sol;p>&NewLine;