.svg)
This is MISSION, a weekly newsletter for people who want to build what's next.
THIS WEEK IN AI
Kevin Roose of The New York Times put his P(doom) at 5%. Meaning, he’s betting there’s a one in twenty chance that AI turns on us and does catastrophic harm.
That lines up with with Matt Clifford, an AI advisor to the British PM, who said, “If we go back to things like the bio weapons or cyber [attacks], you can have really very dangerous threats to humans that could kill many humans – not all humans – simply from where we would expect models to be in two years’ time.”
Ajeya Cotra, an AI safety expert, put hers at 20 to 30%.
Marc Andreesen would say: Of course she does. As a researcher studying “potential risks from advanced AI,” she’s getting paid to foster AI panic.
So which of the two is being dramatic?
In his latest manifesto, Andreesen not only claims that “AI will save the world,” but that putting the pedal to the metal on AI development is our only shot at beating out China in the battle for global AI dominance.
“X will save the world” is always a dubious construction for a sentence. Dialing down the hyperbole one or two notches, and setting aside the extinction question, Andreesen has at least one good reason for AI optimism.
Take the fear that we all have—engineers, designers, even VCs—that an AI bot will take our jobs. Andreesen claims this fear is based on the “Lump of Labor Fallacy.”
“This fallacy is the incorrect notion that there is a fixed amount of labor to be done in the economy at any given time, and either machines do it or people do it – and if machines do it, there will be no work for people to do.”
Instead, he says, we’re going to see a massive increase in productivity, meaning lower prices, greater buying power, higher demand, and real growth leading to a larger economy.
The gist of his take? It’s time to take the AI training wheels off.
If you’re curious about what the future holds for generative AI, tune in on June 15th.
CHART OF THE WEEK
Will Your AI Product Survive?
Rude awakening: 85% of enterprise big data and AI projects fail.
With new AI products coming out left and right, a question hangs in the air: How can savvy businesses tap into the power of GPT in ways that are actually useful to people? Instead of just jumping on the AI bandwagon.
A recent analysis from Reforge shows there’s a delicate balance between the human importance of a challenge and the amount of context necessary for a model to solve it.
On the Y-Axis is “Consideration,” meaning the amount of effort required to make a decision. Buying a car is a high consideration process. Picking a dish detergent requires, for most people, much less.
On the X-Axis is “Context”—that’s the metric for how many abstract concepts a model needs to understand in order to provide a helpful response.
Reforge explains the curve:
“StitchFix: Their model uses AI to augment humans to help you find the perfect outfits. This requires a lot of thought (high consideration), but the options are limited to StichFix’s fashion vertical, rather than an entire shopping mall (low context).
Walmart Text to Shop: This tool helps users automate household shopping. The stakes aren’t quite as high here, but it does require more context with Walmart’s vast e-commerce and in-store assortment of 1.5+ million SKUs.
NotionAI: This GPT3-powered tool requires more training than anything else — the whole internet! However, the stakes are low; you can always brainstorm and write on your own, and there aren’t any compelling alternatives.
Fully self-driving vehicles: This is way above the curve in the top right quadrant because it requires an ungodly amount of training, combined with high stakes to be “right” 100% of the time.
Amazon Alexa: This falls below the curve because most of the use cases are low stakes (checking the weather, turning down the volume) and low context (it can order Domino’s Pizza, but it can’t order you a pie from the best deep dish spot in Seattle).”
A.TEAM EVENTS
Get Your Own AI Hackathon Team
Ever wish you had your own expert generative AI team to take your moonshot ideas to life?
In May, we partnered with OpenAI, Cooley LLP, and Baseten for a transformative Generative AI Enterprise Hackathon, where six A.Team’s paired with Fortune 500 executives to develop groundbreaking solutions for enterprise-level problems — and in just 48 hours they built six functional prototypes.
But we were just getting started.
This June, A.Team is hosting its second generative AI hackathon in partnership with OpenAI and Baseten, and applications are now open to join as an Enterprise Innovation Advisor.
Advisors will be paired up with each hackathon team to brainstorm product ideas and use cases in preparation for June 24-25, when teams will spend the weekend building cutting edge prototypes to explore the wild ways generative AI can revolutionize enterprise.
The winning team will then be invited to demo live on July 14 at Tech Chicago Week—the SXSW of the midwest—to over 5,000 founders, executives, and investors on the Navy Pier Main Stage.
Apply to Advise a Hackathon Team
MISSION MUST-READS
- Don't Buy AI Products Off the Shelf—Build Your Own
- How to Build an AI Enterprise Prototype in a Weekend
PARTING MEME
.jpeg)
.jpg)
