1. Winning strategies for applied AI companies
|
This is a deep dive into the "third wave" of AI startups. Louis Coppey, an associate at Point Nine Capital, creates a framework that categorizes AI companies and looks at how each of the four categorizations affects risks, rewards, and fundraising needs.
+ Coppey's "third wave" is applied AI. Want to learn more about applied AI? Check out the AI Conference in San Francisco (Sept 17–20)
|
2. AI in the software engineering workflow
|
| The workflow of the AI researcher has traditionally been quite different from the workflow of the software developer. Peter Norvig explores how the two can come together.
|
3. A reality check for IBM's AI ambitions
|
Overhyped, IBM's Watson hit a rough patch of PR this year. "Watson is a joke," Chamath Palihapitiya, founder of VC firm Social Capital, said on CNBC in May. But did IBM's inflated expectations and unrealistic timelines knock Watson out of the running, or is there still hope for Watson to improve health outcomes and lower costs?
+ IBM is telling Congress not to fear the rise of an AI overlord
++ A curated list of medical data for machine learning (GitHub)
|
4. Do we think like machines, or do they think like us?
|
| DeepMind researchers gave deep neural networks for image classification never-before-seen images and found that the networks have a tendency to categorize objects according to shape rather than color. Cognitive psychology experiments show humans do the same.
|
5. Bad news for the corrupt
|
| "The [UK's] Serious Fraud Office had a problem. Its investigation into corruption at Rolls-Royce was inching towards a conclusion, but four years of digging had produced a massive pile of documents: over 30 million, including everything from spreadsheets to emails about staff away days." You see it coming, right? Lawyers were sifting through an impressive 3,000 documents a day, but AI processed 600,000 documents a day, at a cost of £50,000—and with fewer errors."
|
 If you follow us on Twitter
#OReillyAI is now @TheAIConf. You might want to follow us there. |
6. 25 AI terms you need to know
|
| Sarah Davis lists definitions of 25 common AI terms.
|
7. Mind reading with machine learning and fMRI
|
| Combining machine learning algorithms with fMRI brain imaging technology, Carnegie Mellon University researchers created a model that can finish your sentences for you with 87% accuracy.
|
8. No investors for AI startups?
|
| A few years ago, Ben Lorica, O'Reilly's chief data scientist, asked, "A decade from now, will we be saying that all businesses are AI businesses?" Apparently, it didn't take a decade. Frank Chen, a partner at Andreessen Horowitz, says "in 2 years, no investor is going to be explicitly looking to fund AI-powered startups...because investors will assume the startup is using the best available AI techniques to solve the problem they are solving."
|
Startup Showcase deadline is midnight Friday, July 21.
|
|
|
9. Cars that coordinate with people
|
| In this excerpt from her AI Conference keynote, Anca Dragan introduces a mathematical formulation that accounts for cars responding to people and people responding to cars. (If you're a Safari member, watch the whole thing here. If you're not a Safari member, try a free 10-day membership.)
|
10. 7 myths about AI that hold your business back
|
| VentureBeat lists seven ways you may be missing the boat on AI.
|
