Profile cover photo
Profile photo
Ethan Hays
790 followers -
Head of Performance Marketing at gyro, web nerd and grilled meat enthusiast.
Head of Performance Marketing at gyro, web nerd and grilled meat enthusiast.

790 followers
About
Ethan's posts

Post has attachment
I love this interview with x.ai founder +Dennis R. Mortensen. It touches on some of the simple truths around the importance of listening to what your customers truly need, even if they won't say it directly ;)

"What entrepreneurs usually do is they hire an office manager, which is a personal assistant in disguise. They just can’t get themselves to say the words ‘personal assistant...

“The funny thing is that everybody, given the opportunity and a $50,000 check, would take that human personal assistant in an instant. But, without that check, it’s cost prohibitive for most people to put that person in place. The interesting thing is that this behavior suggests that there already is a product-market fit. There’s something out there that people want, it’s just too expensive. That, to a large degree, was my catalyst.”


Post has attachment
Slack, I'm Breaking Up With You

"It’s like getting caught in one of those support chats from hell with a Comcast rep who’s clearly trying to simultaneously jockey a dozen text conversations like some kind of bargain basement Bobby Fischer, except that it’s all day long and with everyone I know."



Post has attachment
“The connections between your Shatner’s Bassoon and the peri-sinovial labio-ventral cortex is unusually cromulent and this might predispose you to brilligness, and excessive slithy toveing.”

OMG do I ever love this site. As a scientist, it pains me greatly to see pseudo-science practitioners profiting off of the public, who haven't been educated on how to read scientific research critically.

The author of NeuroBollocks can read scientific research critically, and can write about it hilariously.

#cromulent #woo #pseudoscience 

Post has attachment
Iran's Blogfather Discusses the Internet after 8 Years of Prison

Fascinating perspective from someone who was deeply involved in web publishing from 2001 - 2008, jailed 2008 - 2016, and now free.

He makes really deep, substantive points about how platforms (Facebook, Instagram, Twitter, etc) and "The Stream" have changed what the web is:

"But hyperlinks aren’t just the skeleton of the web: they are its eyes, a path to its soul. And a blind webpage, one without hyperlinks, can’t look or gaze at another webpage – and this has serious consequences for the dynamics of power on the web.

We seem to have gone from a non-linear mode of communication – nodes and networks and links – toward one that is linear, passive, programmed and inward-looking.

Maybe this is all a natural evolution of a technology. But I can’t close my eyes to what’s happening: a loss of intellectual power and diversity. In the past, the web was powerful and serious enough to land me in jail. Today it feels like little more than entertainment."

Post has attachment
2016 will be a fascinating year in Artificial Intelligence. Effectively, you have different sets of competitors trying to seed the market in their favor by open-sourcing the stuff that's the least proprietary (or advantageous) to their own companies:

1) TensorFlow: Google's ML algo is probably the least proprietary part of the ML stack (the other components being massive data and massive infrastructure)

2) Open Compute Project: Facebook open-sourced the problem of building datacenters at scale (where they are at a disadvantage to Google)

3) OpenAI: Tesla & YCombinator commit $1B to developing open-source AI (where they are at a disadvantage to Google & FB)

Each open source project attempts to develop the market in their favor. To take the TensorFlow example:

* They're leveraging the collective knowledge of ML experts all over the world to make TensorFlow better.
* In turn, this ensures that the ML experts of the future grow up with TensorFlow as a default area of knowledge.
* Finally, this cements TensorFlow as one of the world's default ML systems, which (conveniently) utilizes standards set in place by Google itself, with a design already suited for Google’s infrastructure.

Each initiative follows much the same logic (at least to my mind). I think 2016 is going to be a bonkers year for AI & ML, and I kinda can't wait to see where it goes....

#ai #artificialintelligence #machinelearning #google #rankbrain

Post has attachment
Good Product Manager / Bad Product Manager

Calling out the qualities of a bad PM at first seemed a bit negative, but it really made me reflect. Good read http://bit.ly/1l4Ze3X

#productmgmt #pm

Photo

How Not to Write About Measuring Ad Campaigns

"Advertising is not an exact science. There’s no precise way to measure the success of an ad campaign."

The title of the article is "Metrics for Measuring Ad Campaign Effectiveness" and those are the first 2 sentences.

Yuck.  I already knew that AllBusiness.com was a generally pretty shitty in the content-farm, written-by-someone-with-passing-knowledge way, but I didn't realize it was that bad....

#measurement #analytics  

Post has attachment
We May Already Be at Our SkyNet Moment

+Tim O'Reilly has a really fresh view of the whole "singularity" issue:   http://bit.ly/1LWg0L7

"I believe that if you understand that an AI might not be an independent intelligence but rather a system that includes us -- all of us -- as a component, we may already be at our Skynet moment.

"What we call "The Market" is already hostile to humans, in many ways. When companies say they have to outsource manufacturing to sweatshops, or Southwest Airlines says they have to start charging baggage fees because otherwise "the market" will punish them, what they are saying is that they are embedded in a system that we no longer control, as Wallace Stevens said (speaking of something else), "without human meaning, without human feeling, a foreign song." So many of our systems are already machines that rule us."

#ai #machinelearning  
Photo

Post has attachment
Should You Ever Use a Pie Chart?

No.

Well, actually that's a bit simplistic.  If you only have 3-4 data categories, it's probably fine, and pretty intuitive in that it conveys that all elements have to add up to 100%.  But pie charts tend to be misused.

#dataviz
Wait while more posts are being loaded