The LAST Unusual Loop convention is correct across the nook!
The convention has collected 673 movies on its YouTube channel, which implies there’s gold in them hills, however discovering the gold would possibly devour a few of your treasured time. Excellent news!
I used to be hanging out within the convention’s Slack when somebody requested:
What are a few of y’all’s favourite previous Unusual Loop talks? As somebody who is aware of the conf by fame however hasn’t been earlier than, I’m curious what a few of your favorites are!
A barrage of YouTube hyperlinks adopted, so I scooped them as much as share right here so everybody can get pleasure from these golden Unusual Loop talks ✨
🎥 Cease Writing Lifeless Packages by Jack Rusher (2022)
Most new programming languages are unintentionally designed to be backwards appropriate with punchcards. This speak argues that it could be higher to deal with constructing new stay programming environments that may assist us remedy the issues of the longer term.
Josh says: “That is the speak that satisfied me ‘if this can be a matter mentioned at Unusual Loop, then I really want to make it there’”
🎥 Easy Made Simple by Wealthy Hickey (2011)
Wealthy Hickey, the writer of Clojure and designer of Datomic, is a software program developer with over 30 years of expertise in numerous domains. Wealthy has labored on scheduling techniques, broadcast automation, audio evaluation and fingerprinting, database design, yield administration, exit ballot techniques, and machine listening, in quite a lot of languages.
This keynote was given at Unusual Loop 2011, and is maybe one of the best identified and most extremely regarded of Wealthy’s many wonderful talks, ushering in a brand new approach to consider the issues of software program design and the fixed battle in opposition to complexity.
See additionally: Wealthy Hickey’s biggest hits
🎥 A Entire New World by Gary Bernhardt (2012)
This speak declares probably the most bold software program mission I’ve ever undertaken, then considers why its existence is so stunning (and in some instances irritating) to individuals.
Conrad says: “I’m personally engaged on this: fully tearing down the legacy of punchcards in trendy dev tooling.”
🎥 Languages for 3D Industrial Knitting by Lea Albaugh (2016)
Industrial computer-controlled knitting machines are churning out seamless three-dimensional material objects as you learn this – assume gloves, sweaters, sportswear, furnishings, architectural materials, and plushies. However proper now the programming languages for these machines are horrible.
With a knitting machine, advanced buildings are shaped with a really restricted set of low-level machine operations. Considering on the extent of these machine operations turns into untenable fairly shortly once you begin making an attempt to make actual objects. How can we guarantee knittability? Can we encode constraints in an comprehensible approach? Can we optimize manufacturing velocity?
We’ll have a look at a number of methods to explain knit buildings and generate machine directions for his or her fabrication, together with each the delightfully arcane language that ships with the machine and the language from our SIGGRAPH 2016 technical paper, “A Compiler for 3D Machine Knitting.”
🎥 African Polyphony & Polyrhythm by Chris Ford (2016)
Ethnomusicologists face a dilemma: both shoehorn African music into European notation, or create customized DSLs that may solely be understood by a choose band of European ethnomusicologists. Algomusicologists can remedy this drawback, as a result of we now have principled methods of modelling particular idioms when it comes to common computation. What’s extra, our fashions could be executed to generate precise music.
Simha Arom is a French-Israeli ethnomusicologist. Within the guide from which the title of this speak is borrowed, he describes the rules underlying the musical system of conventional central African polyphony and polyrhythm. Arom invented ingenious recording strategies for deconstructing and systematising musical cultures that had no earlier custom of musical principle. He examined his fashions by utilizing them to recreate music and alluring central African musicians to critique the outcomes.
This speak will argue that music-as-code is a perfect strategy to symbolize Arom’s insights. I’ll present the complete fact within the Lévi-Strauss quote with which Arom prefaces his guide: “The proof of the evaluation is within the synthesis.” Programming has performed its half within the homogenisation of world tradition, however the universality of the lambda calculus additionally affords a golden alternative for code to turn into a degree of interchange between previously incompatible musical techniques.
Brian says: “My two favourite talks are in all probability African Polyphony & Polyrhythm by Chris Ford and Easy methods to Hack a Portray by Tyler Hobbs. But it surely’s arduous to decide on, there have been so many wonderful ones through the years…”
🎥 Easy methods to Hack a Portray by Tyler Hobbs (2017)
In Hackers and Painters, Paul Graham famously linked coding and the visible arts. On this speak, I’ll clarify do each without delay.
By means of the Processing framework, Clojure, and cautious remark, it’s potential to simulate watercolor paints and different media. I’ll stroll the viewers via the remark of actual watercolor paint, breaking down the psychological mannequin we are able to use to consider its form, texture, and qualities. Subsequent, we’ll translate that psychological mannequin right into a Clojure program able to producing watercolor results. Particular consideration can be paid to chance distributions, as randomness is the important thing to pleasing results.
With the fundamentals established, we are able to dive into textures and the mixing of a number of colours. Lastly, I’ll reveal mix every part to create a completed generative murals, able to producing a stream of distinctive however aesthetically associated pictures.
🎥 Easy methods to Repair AI: Options to ML Bias (And Why They Don’t Matter) by Joyce Xu (2019)
Bias in machine studying is a Drawback. That is frequent data for many people now, and but our algorithms proceed to function unfairly in the actual world, perpetuating structural inequality alongside strains of sophistication and colour. In spite of everything, “higher coaching knowledge” is just not really easy to get our palms on, proper?
On this speak, I argue that it’s time for us to start constructing algorithms which might be designed to be resilient to biased knowledge. Constructing on a fundamental introduction to ML ideas, I current an in-depth, intuitive rationalization of a number of deep studying strategies that fight underlying bias in knowledge, and use these fashions to discover what “algorithmic equity” actually means in measurable phrases. Lastly, diving into just a few case research of actual world techniques, I counsel that even good “equity” is just not essentially the fairy-tale ending we prefer to assume it’s. Blindly optimizing for it might nonetheless miss the actual drawback behind AI bias, and to return to an actual answer we could need to reframe the issue itself.
🎥 Cease Price Limiting! Capability Administration Achieved Proper by Jon Moore (2017)
In an period of cloud computing and microservices, it’s crucial to have the ability to isolate shoppers from each other when it comes to the sources they’ll use–in any other case, one misconfigured or malicious shopper could make a service unusable for everybody else. Steve Yegge mentions quotas and throttling as key practices in his infamous “platform rant.”
For those who’ve ever had one in every of your inner microservices launch an unintentional denial of service assault in opposition to one other one, this speak is for you!
🎥 Misuser by David Schmudde (2018)
The act of discovering and displaying digital artwork created for twentieth century computing and telecommunication environments locations the character of those techniques in stark aid. Most exhibitions require an archeological endeavor and an unlimited effort to execute a single artifact from forty years in the past.
The software program I characteristic on this speak operated on the fringe of digital techniques. Artists typically behaved like hackers: working off hours, pushing a system past its specification, and inventing new potentialities for human interplay alongside the best way. They have been a few of the most keen practitioners of early computing however typically discovered it troublesome to acquire entry. Their work typically noticed restricted distribution and few updates.
These challenges exasperate the problem of constructing a retrospective of digital artwork. The software program’s brittle nature displays the surroundings inside which it ran. The engineering selections that result in these techniques have been typically knowledgeable as a lot by tradition as science. “Misuer” will illustrate this connection and construct a framework to look at the worth of conventions and what we are able to be taught by working outdoors of them.
Daniel says: “This one made me a little bit unhappy, in a great way”
🎥 Mapping Imaginary Cities by Mouse Reeve (2018)
Whereas the map is just not the territory (to cite the semantician Alfred Korzybski), the map remains to be normally supposed to correspond to at least one. However what about maps of nowhere in any respect? What can they symbolize and the way can they be made?
Maps are a well-known a part of each day life, with a deeply acquainted and complicated symbolic language, and an extended historical past. They’re additionally vastly diversified in fashion and aesthetic, and infrequently are artistic endeavors unto themselves. All this makes mapping a strong inventive software for conveying concepts a few area, how it’s used, and who inhabits it. But it surely additionally presents a mapmaker with what can really feel like an amazing array of design selections and technical hurdles to beat to be able to create a generative map.
This speak will discover maps as a strategy to talk about individuals and place within the context of fictional cities, and dive into algorithms and strategies for procedurally producing maps by increase topography, panorama, populations, and road plans.
🎥 The Mess We’re In by Joe Armstrong (2014)
Joe Armstrong is among the inventors of Erlang. When on the Ericsson pc science lab in 1986, he was a part of the crew who designed and applied the primary model of Erlang. He has written a number of Erlang books together with Programming Erlang Software program for a Concurrent World. Joe has a PhD in pc science from the Royal Institute of Expertise in Stockholm, Sweden.
🎥 Vellvm – Verifying the LLVM by Steve Zdancewic (2018)
LLVM is an industrial-strength compiler that’s used for every part from day-to-day iOS growth (in Swift) to pie-in-the-sky tutorial analysis tasks. This makes the LLVM framework a candy spot for bug-finding and verification applied sciences–any enhancements to it are amplified throughout its many purposes.
This speak asks the query: what does LLVM code imply, and, how can we make sure that LLVM-based instruments (compilers, optimizers, code instrumentation passes, and so forth.) do what they’re purported to – particularly for security or safety essential purposes? The Verified LLVM mission (Vellvm) is our try to supply a solution.
🎥 Programming Ought to Eat Itself by Nada Amin (2014)
Nada Amin is a member of the Scala crew at EPFL, the place she research sort techniques and hacks on programming languages. She has contributed to Clojure’s core.logic and Google’s Closure compiler. She’s liked serving to others be taught to program ever since tutoring SICP as an undergraduate lab assistant at MIT.
Michael says: “This speak nonetheless blows my thoughts a little bit (as do most talks by Nada Amin)”
🎥 The Early Days of id Software program: Programming Rules by John Romero (2022)
As co-founders of id Software program, John Romero and John Carmack created the code behind the corporate’s seminal titles. The rules they outlined via expertise in id’s earliest days constructed upon each other to provide a novel methodology and a always shippable codebase. On this speak, John Romero discusses id software program’s early days, these programming rules and the occasions and video games that led to their creation.
Larry says: “As somebody who grew up enjoying these video games, it was sort of thoughts blowing listening to how they have been developed. I believed the emphasis on investing in tooling was actually fascinating… and he’s an incredible storyteller :)”
🎥 Easy methods to Take into consideration Parallel Programming: Not! Man L. Steele Jr (2010)
Man L. Steele Jr. believes that it shouldn’t be the programmer’s job to consider parallelism, however languages ought to present methods to transparently run duties in parallel. This requires a brand new strategy in constructing languages supporting algorithms constructed on independence and build-and-conquer rules somewhat than on linear decomposition of issues.
Steven says: “This keynote the place Man Steele dissects an outdated punch card program he wrote left my jaw on the ground… and nonetheless does any time I rewatch it”
🎥 Add ALL the issues: summary algebra meets analytics by Avi Bryant (2013)
A mathematician buddy of mine tweeted that anybody who doesn’t perceive abelian teams shouldn’t construct analytics techniques. I’d flip that round and say that anybody who builds analytics techniques finally ends up understanding abelian teams, whether or not they comprehend it or not. I’ll discuss how the legal guidelines of group principle present a helpful codification of the sensible classes of constructing environment friendly distributed and realtime aggregation techniques, and provides some examples of how that maps to helpful code.
I’m certain there are many extra nice Unusual Loop talks that didn’t shared by this specific group of attendees! Do you might have a favourite that didn’t make the checklist? Remark beneath, and we hope to see you at Unusual Loop 2023! 💚
