September 2015 Newsletter September 14, 2015 | Rob Bensinger | Newsletters Research updates New analyses: When AI Accelerates AI; Powerful Planners, Not Sentient Software New at AI Impacts: Research Bounties; AI Timelines and Strategies New at IAFF: Uniform Coherence 2; The Two-Update Problem Andrew Critch, a CFAR cofounder, mathematician, and former Jane Street trader, joined MIRI as our fifth research fellow this month! As a result of our successful fundraiser and summer workshop series, I’m happy to announce that we’re hiring two additional full-time researchers later in 2015: Mihály Bárász and Scott Garrabrant! General updates We’ve wrapped up our largest fundraiser ever: 258 donors brought in a total of $629,123! Thanks to you, we’ve hit our first two fundraising goals and received $129,123 that will go toward our third goal: Taking MIRI to the Next Level. “If you agree AI matters, why MIRI?” Two new replies to this question: Assessing Our Past and Potential Impact and What Sets MIRI Apart? We attended the Effective Altruism Global conference. Background: AI and Effective Altruism. We’re moving to a larger office in the same building this October. If you have experience in office planning or access control systems and would like to help out with our plans, shoot Malo an email at firstname.lastname@example.org. News and links GiveWell publishes its initial review of potential risks from advanced artificial intelligence, as well as a new article on the long-term significance of reducing global catastrophic risks. Stuart Russell explains the AI alignment problem in a San Francisco talk and in a (paywalled) interview in Science. A much earlier discussion of the AI control problem in a 1960 Science article. From Vox: Robots Aren’t Taking Your Jobs — And That’s the Problem.