Jamais Cascio: “I Can Has Singularity?”, the Need for Precision in AI Reporting

Jamais Cascio has a good article up at Fast Company on IBM’s “cat brain” project. My thoughts are the same as his. A cat-sized neural simulation is one thing, an actual cat emulation is another thing entirely. IBM is branding its achievement as if it made Mr. Fluffles (Uploaded EditionTM) already, with all the cat pictures on their web site, but this is all a bunch of hype. Mr. Cascio’s title is actually misleading because it says “IBM Simulates the Inner Workings of a Cat Brain”, later noted in the article itself as false.

Blue Matter is another interesting project, but another non-game-changer. Read the Whole Brain Emulation Roadmap for a bigger, hype-free, never-touched-by-the-marketing-department picture of the whole field.

It is extremely easy to give this story incorrect connotations. For instance, over at Futurismic, Paul Raven, who usually does quite careful and reliable reporting, is using some slightly misleading language that is beyond what is found in the press release:

But delivery systems aside, what’s the story here? Basically, IBM have built a computer that simulates the complexity and interconnection of a cat’s brain, which is significantly more complex than previous neuro-cortical simulations.

Yes, IBM has built a computer that simulates a cat-SCALE brain, one that has complex features, but these complex features are nowhere near as useful for intelligence as the complex features that allow a cat to be clever. It doesn’t even simulate the features of intelligence that allow cockroaches or honeybees to be clever. It is just a proof of concept for the canvas for such simulations. It’s like the difference between building the ceiling of the Sistine Chapel and painting it. It’s certainly nice, yes, if you like that sort of thing, but I think we should keep the hype machine set to “low” for the time being.

AI as a field is progressing, but many AI-related press releases you see on the Internet do not really describe major progress forward. Some of the more interesting progress is happening behind the scenes with non-press-release-friendly mathematics and theoretical work, including computational cognitive science.

Comments

  1. Y’know, as someone who writes copy for a living, I should really be better at spotting obfuscation in other people’s… my failing is that I hear what I want to hear sometimes. Consider this a lesson learned – I’ll wait for detailed analysis of claims in future. :)

  2. red

    I think it’s okay to celebrate small milestones with regard to the important work being done there, but your right, it can be too easy to get swept up in the hooey. If the AI community doesn’t slap down this kind of nonsense hard and fast it’s only going to make it that much harder for the mainstream media to take it seriously when something truly impressive happens.

Leave a Comment

Awesome! You've decided to leave a comment. Please keep in mind that comments are moderated.

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>