Are Investors Misinterpreting arXiv As Peer Review?

A deep-tech investing warning: founders increasingly cite arXiv preprints in funding announcements, and investors under deal pressure read published on arXiv as peer reviewed. It isn’t. The question the piece leaves standing is whether partners are making quantum/AI investment decisions on the strength of papers nobody has refereed.

Notes

• We’re currently tracking VC firms across Europe & the US that are investing in emerging deep-tech such as; quantum computing, artificial intelligence and distributed systems. We’ve consistently noticed partners at VC firms being very excited about portfolio founders publishing papers on the arXiv pre-print service. Of course, there is nothing wrong with partners promoting the work of their founders! :)
• Moreover, we’ve also noticed a sizable number of founders referencing arXiv papers in their funding announcements and marketing material. This seems to indicate that publishing on arXiv before fundraising carries significant weight. Hence, we couldn’t resist asking ourselves the following:
  • • Could partners at VC firms, under pressure to close what appears to be a hot deal, be making deep-tech investment decisions based on founders’ papers being published on the arXiv?
  • • Could partners with a non-academic background even potentially be misinterpreting arXiv as peer review?
• To help us uncover the answers to the questions above, our team at Kähler Insights is recruiting investors (Angels, VCs, PEs) willing to schedule a 25 minutes interview to discuss the impact of arXiv on their investment decisions regarding emerging deep-tech fields i.e. quantum computing. If you are an investor and would like to participate, please feel free to fill out this form. We’re particularly keen on pre-seed and seed type of investors given that deep-tech founders are often just spinning out of research labs before closing their first rounds!
• What is arXiv anyway?
• In a nutshell, papers are submitted to arXiv to enable authors to rapidly and openly share their work with the wider research community. Although papers go through qualified moderators for QA before being accepted, arXiv is not peer review because the moderators are generally not deep experts nor do they have the time to rigorously analyse papers. Not surprisingly, papers that get accepted in arXiv don’t necessarily pass the rigorous peer review process of journals! In the publications ‘How many preprints have actually been printed and why: a case study of computer science preprints on arXiv’ & ‘Popularity of arXiv.org within Computer Science’- the following interesting data (2008 – 2017) can be extracted:
  • • Usage of arXiv.org has risen dramatically among the most selective conferences in computer science. In 2017, 23% of papers had e-prints on arXiv.org, compared to only 1% of papers ten years before.
  • • Areas of computer science vary widely in e-print prevalence. In theoretical computer science and machine learning, over 60% of published papers have arXiv e-prints. In other areas, arXiv usage is essentially zero. In most areas, arXiv usage is rising.
  • • Of the 2017 published papers with arXiv e-prints, 56% were preprints that were posted either before submission or while the review process was ongoing.
• Should investors trust ArXiv?
• The short answer is yes, but only as part of a qualified lead pipeline, not as the absolute factor for writing a check! The truth is, emerging deep-tech fields such as quantum computing are still fairly inaccessible for those investors with a non-academic background as specific academic experience enables one to appreciate the nuance in technical papers & helps one to differentiate between small incremental or trivial technical progress versus a real breakthrough. Hence, as deep-tech investor, never fall for a deal just because the founder(s) wrote a paper published on arXiv. Likewise, don’t let yourself be fooled by a founder’s large number of publications on arXiv - as the old say goes ‘it’s quality not quantity’!