Brian Dennett

Any archiving services tracking @GooglePlay reviews? I don't think @Google did anything nefarious, an anti-spam system just kicked in for @RobinhoodApp . However, a lot of those reviews were legitimate and I'd love to get my hands on them. @internetarchive perhaps?

Yesterday I was telling people the wallstreetbets frenzy would be a historical footnote. Now it's starting to look like it will deserve its own chapter. Did Robinhood just become a sacrificial lamb so larger players could get out?

In October of 2008, Volkswagon was briefly the most valuable company in the world. If you don't know the story, it's a fascinating one. Gamestop (GME) has been experiencing a similar situation lately. I'm not invested but I am most certainly entertained. ft.com/content/0a58b6…

@robmay If social networks were better stewards of what was shared when the subject matter wasn't gray, how much longer might it have been before these conversations were warranted? Censorship is a hard problem but the complete lack of it has created unnecessary divisiveness.

@robmay Exactly the crux of the issue. It's not solely about censorship, it's about dissemination and promotion. Social networks are highly opinionated on what gets shown, that is the point at where a higher standard should be held.

Excited to be part of the Awesome Inc fellowship program.

Cool to see @diginomica with the #1 article on Hacker News right now, you've gotten the attention of the techies with this one news.ycombinator.com/item?id=216139… @krmarko @jonerp @dahowlett

Is shared powerpoint templates really still an unsolved problem in 2019? Feels like I'm wading through a cesspool of dark patterns trying to find a decent option.

@grahamrobbo Blasphemy

Just uninstalled Sublime - I am now all in with vscode

@dahowlett "Essentially unchanged" may be mostly defensible from a math/theory perspective but practically speaking the field has made plenty of progress in plenty of areas. And by field I mean ML/DL, AI is an ambiguous enough term that it tends to not be helpful when talking progress.

Great resource for browsing ML research paperswithcode.com

re: eng.uber.com/aresdb/ are GPU accelerated DBs about to have their moment in the spotlight? Who's going to bring this to the Enterprise? @jonerp @holgermu @rhirsch @josheac

.@robmay You should check out the latest dlvr.it/QcdyxQ from @jonerp, it overlaps a lot with your commentary on NLP from 7/15. The guys over at @diginomica have been having an ongoing discussion about AI and CX in the Enterprise that I think you'd appreciate.

@jonerp @philww "most impactful applications of AI are those where people aren’t even aware of the technology" sounds a lot like the AI effect en.wikipedia.org/wiki/AI_effect

"Cloud Next 18 – How Google sees the future of AI in the enterprise" diginomica.com/2018/07/24/clo… -> As #GoogleNext18 kicks off, @philww shares an AI preview with Google’s cloud AI and machine learning team lead. Unstructured data is one big theme...

@jonerp @dahowlett I commend British Airway for showing everyone how to learn GDP the hard way techcrunch.com/2018/07/19/bri…

@Steve_Lockstep @grahamrobbo @timoelliott @TomRaftery I didn't write the rules. Arguably it's a matter of good faith that businesses in GDPR compliance aren't simply obfuscating their user tracking but avoiding the effort of individual tracking all together and taking appropriate measures to remove the temptation to do so.

@Brian_Dennett @grahamrobbo @timoelliott @TomRaftery Why does regulatory compliance soften the standard expected of deidentification?

Data Ownership: The Next Big Data Threat and Opportunity timoelliott.com/blog/2018/07/d… < I'd be very interested to hear any feedback from others. This topic seems very under-debated to me... #analytics #infonomics

@Steve_Lockstep @grahamrobbo @timoelliott @TomRaftery I don't disagree but we're talking in the context of regulatory compliance. For adversarial situations, security via obscurity is never a good practice and anonymization is exactly that.

@grahamrobbo @Brian_Dennett @timoelliott @TomRaftery Most privacy and security experts believe it is no longer possible to keep any promise of data anonymization. IMO privacy needs to be protected with a multilayered defense-in-depth approach. Deidentification is not enough.

Brush up on the underlying math or find choice bits from the latest academic paper? The never-ending question of the machine learning practitioner.

@jonerp @ewH I realized we never circled back to the engineering rigor, not to mention scientific rigor, required to do ML well in practice. Definitely a lot of ground to still be covered on the topic of "practical AI"

"Practical AI in focus - a chat on customer use cases and AI hype overload with @Brian_Dennett of Enable AI" -> enjoyed this podcast taping at #sapphirenow, jonerp.podbean.com/e/practical-ai… … (not SAPpy though content-wise) cc: @ewH