Slack CEO Stewart Butterfield:
"the opportunity we see together is massive … as software plays a more and more critical role in the performance of every organization, we share a vision of reduced complexity, increased power and flexibility, and ultimately a greater degree of alignment and organizational agility. Personally, I believe this is the most strategic combination in the history of software, and I can't wait to get going."
It seems he couldn’t quite figure out how to work the words “synergy” and “nimble” in there.1😂I doubt it, given that I saw that only 1 of 8 cores was busy during the times I was watching. I am nearly certain it did go to sleep for 30 to 60 minutes somewhere in the middle, though, based upon some modification times on different output files.
sounds like an ai wrote that :)
1😆waiting room is on by default now, but you can change that back
a lot of the new default settings are designed to be better/safe for teachers etc
zoom has often been better in that calls usually worked
Ok, this is legit funny. https://twitter.com/slack/status/1333886283689709569
Surely by now they have offered him a not-insubstantial sum for that handle
any general opinions on using Clojure foreign function capability (interoperable) with Python, as well as Clojure Script?
I am trying to think of the trade offs between directly calling foreign code… versus using a message broker like Kafka
process boundaries are a fine place to build a stronger abstraction
don't need kafka for that
shell invocation, http, all kinds of ways to handle that, some of which are slower and simpler, some are faster.. some sync vs async
in short, I think you will spend as much time trying to call a python runtime properly from JVM
as you would building a cleaner interface between processes, or subprocesses
ahh ok. thank you so much for that update…
I would never ship clojure in jvm calling python runtime BTW
you have to go thru clojure -> java -> JNI -> Python runtime ...
so many abstractions and translations, best to just output spit and slurp some JSON over a socket or stdin 8^)
OTOH https://github.com/clj-python/libpython-clj is well-supported and fairly heavily used by the data science folks -- it's grown out of that community.
1😁The same folks have just released a Julia/Clojure interop library BTW.
1😁That said, I would probably only bother going across that sort of boundary if I had a lot of experience in both languages and there were some critical libraries I needed to use in the non-Clojure ecosystem 😐
https://adventofcode.com/ day 2 is starting in 10 min, time to warm a REPL.
5🔥Slack does threads terribly. Just look at this conversation. Try to follow it. Try to notice a new topic introduced in this noise. It’s intentional. The only way to truly make sense of it is to be a part of it in real time, aka “boost engagement.” There’s no way Slack is useful as an archive. Slack is a distraction machine
Twist 🤞
yes, I was looking at the new clojure-python lib… Sounds like I should only experiment with that… not on production… maybe just for fast AI/Data type experiments.
I have used foreign functions on Ruby/Java… and yea… sometimes there’s a hiccup… 1 little incompatible call
THREE
(Look at the picture at the bottom)
• you can choose either side of the path
• you can get to the path from anywhere
• the paths can merge.
THREE
• Choice
• transformation
• join
THREE
(if (nil? %) :success nil)
nil? % => choice
if => transform
:success nil => join
THREE=>MONAD
(macroexpand-1 '(some-> nil inc)) ;; => (let ;; [g nil] ;; (if (nil? g) nil (-> g inc)))
https://youtu.be/YR5WdGrpoug?t=1600 > you would never say "maybe anything" because when your talking about something in isolation, destined to be combined in aggregate, who knows that its maybe? What rich is saying is that the Maybe Monad represents a bad idea. It's possible, but its not something you want to encourage. The issue is that its a choice between complete success and complete failure. try catch success fail. The idea that you might have to makes a choice isn't what we do in programming, we make choices. A useful type is one that helps you understand the choice that was made. If a type just tells you that some choice was a choice, its usefull and its distracting and its pervasive. which is why it becomes necessary to put it everywhere once you adopt the idea. Because everywhere is nowhere at all.
Put another way, intentionally trying returning maybe is like intentionally returning nil.
I should note some of the literature suggests monads are associative. Which would mean your if expressions would need to be pure. i think
I don't see that in the category theory versions, I'm not sure they consider the idea of order at all
@timok fun / scary version: analyze interactions between managers and employees to train an AI middle manager :D
1criminal version: train a model to correlate slack activity and stock price of that company
1😎end to end encryption would be possible in a group messaging scenario, but transcript maintenance and search is in tension with those goals
Salesforce Slack will start doing the same stuff Microsoft was doing last week with the "engagement metrics"
the (global) security implications of Slack are terrifying to me without the Salesforce acquisiton
3💯The Major gazed upon his land A might maze, his people lost, his job to make a plan! What to do? Where do they need to go! His advisors worked late into the night The strained with all their might A plan to let everyone know where to go When morning came, what is to be our fate? The city roads, now blocked by gates! The major his a head a thunder What is this massive blunder! The advisors, aghast, do protest Your excellence, the problem is addressed See now your subjects are never lost For they cant go adrift The major, his eyes a dark and stormy sea There is a problem, dont you see! for there is only one key!
@paul.legato i love your version. it inspired me to make my own 🙂.
Irrational valuation of collaboration technology which is seen as table stakes for SalesForce to play
in the "converged collab" space
aka, docs, storage, wiki, internal resource mgmt, comms etc...
maybe I'm just jaded watching middling exec teams try to buy their way into a story about the next big market trend
given that collaborative tools become the source of truth about a company's workflows and internal knowledge, I don't think it's undervalued, perhaps just not exploited yet
the "rational" thing would be to make it as cheap, simple, and flexible as possible, until it's impossible to migrate out of, then the terms are changed
which sounds like something salesforce would do, to me
at the startups where I've worked, I'd estimate that losing access to data in salesforce (not to mention the salesforce specific workflows) would be more disruptive than losing 3/4 of the sales team