platform fatigue

With the myriad of platforms used to counteract the absence of most direct interactions at work, I start to suffer from platform fatigue, constantly switching to a different interface and wasting a lot of time on retrieving links from old emails and reentering passwords… Even a single platform like Teams requires permanent juggling between Dauphine and Warwick (obviously linked with Microsoft constraints), plus repeated updates that clash more often than enough with Firefox. Not to mention collaborative systems like Overleaf, Wikimath, Git, and others. Same thing for regular Zoom meetings which fail to reopen from one week to the next. And calendars that cannot keep track of everything or even anything! The only interface that keeps working (for me) across accounts is my Thunderbird email interface, except for the sharp increase in the email volume (and the fact that many now bypass emails for chats on Teams, Slack, and another myriad of platforms).

love thy command line [Bourne again]

“Prebuilt into macOS and Unix systems (…) the command line (also called the shell) is a powerful text-based interface in which users issue terse instructions to create, find, sort and manipulate files, all without using the mouse. There are actually several distinct (…) shell systems, among the most popular of which [sic?] is Bash, an acronym for the ‘Bourne again shell’ (a reference to the Bourne shell, which it replaced in 1989).”

An hilarious rediscovery of the joys of shell (line) commands in Nature! Which I use by default for most operations on my computer, albeit far from expertly (despite the use of a cheat tee, from time to time!). One of the arguments in the article, “The mouse doesn’t scale,” is definitely mine as well. Among other marketing lines, wrangling files with no software interference (check), handling huge files (very rarely), manipulating spreadsheets (I don’t), parallelising work on remote servers (check), automate via cron (not anymore)…. Unsurprisingly, most of our students are never using terminals of command lines.

errno EFBIG

ubuntu update

As I had delayed my ugrade of my Ubuntu 16.04 Linux OS way beyond the reasonable, I eventually found myself forced to do it in the least favourable version earlier this week, as the University was closed and my computer support friends were on (well-deserved) vacation. Indeed, my graphical interface went blank and no amount of rebooting was making any change, while I had access to a line editor which was kindly suggesting the command for the upgrade. After a moment of hesitation, I followed the recommendation and prepared for either a long wait or a tragic error message, but nothing like that happened in that the download of the new version was lightning fast and the upgrade itself lasted less than one hour, leaving me with a fully operational 18.04 version. And amazed at recovering my earlier tuning, with even better performances..! (The only glitch is that the colours seem to have turned more reddish or auberginish.) One worry I had with using a Huion graphics tablet [that I had just bought with my amazon credits in preparation for my on-line classes next semester] vanished when I found it had the same minimal performances as earlier [meaning I can write using Gimp, although this app always challenged me, Xournal, which allows for writing on top of my pdf slides, or Microsoft Whiteboard, which works like a potentially infinite whiteboard, but I have no access to the physical control buttons on that tablet]. A happy ending or rather beginning with Bionic Beaver, then, albeit a possibly short one as I should move now to Ubuntu 20.02 version, expecting a similarly smooth transition unless my HP laptop has some compatibility issues…

Laplace’s Demon [coming home!]

A new online seminar is starting this week, called Laplace’s Demon [after too much immersion in His Dark Materials, lately, ather than Unix coding, I first wrote daemon!] and concerned with Bayesian Machine Learning at Scale. Run by Criteo in Paris (hence the Laplace filiation, I presume!). Here is the motivational blurb from their webpage

Machine learning is changing the world we live in at a break neck pace. From image recognition and generation, to the deployment of recommender systems, it seems to be breaking new ground constantly and influencing almost every aspect of our lives. In this seminar series we ask distinguished speakers to comment on what role Bayesian statistics and Bayesian machine learning have in this rapidly changing landscape. Do we need to optimally process information or borrow strength in the big data era? Are philosophical concepts such as coherence and the likelihood principle relevant when you are running a large scale recommender system? Are variational approximations, MCMC or EP appropriate in a production environment? Can I use the propensity score and call myself a Bayesian? How can I elicit a prior over a massive dataset? Is Bayes a reasonable theory of how to be perfect but a hopeless theory of how to be good? Do we need Bayes when we can just A/B test? What combinations of pragmatism and idealism can be used to deploy Bayesian machine learning in a large scale live system? We ask Bayesian believers, Bayesian pragmatists and Bayesian skeptics to comment on all of these subjects and more.

The seminar takes places on the second Wednesday of the month, at 5pm (GMT+2) starting ill-fatedly with myself on ABC-Gibbs this very Wednesday (13 May 2020), followed by Aki Vehtari, John Ormerod, Nicolas Chopin, François Caron, Pierre Latouche, Victor Elvira, Sara Filippi, and Chris Oates. (I think my very first webinar was a presentation at the Deutsche Bank, New York, I gave from CREST videoconference room from 8pm till midnight after my trip was cancelled when the Twin Towers got destroyed, on 07 September 2001…)