aaron

motto

@Vraj247 @raunofreiberg @Gavmn @rsms @ANoffsinger estrella is also very cool amongst his design works

So many people became design engineers after seeing @raunofreiberg and @Gavmn, but my love for design engineering started after getting to know the legendary @rsms — after visiting @ANoffsinger's portfolio and finding the src of rsms.me somewhere in the CSS. If you're a young designer/engineer out of college, you really wanna look at who Rasmus Andersson is!

@sid_srk Sid from Catalyst

@aarnphm A bygone era but Catalyst will take its place, I’m hyped

alphas shawarma is the reason why I moved to Toronto btw

@sid_srk Sid on foundation is number 2 reason btw :)

@aarnphm lfggg

@aaaronwill @avec Haha just my apartment is very dusty (cuz i clean basically everyday)

@aarnphm @avec from one aaron to another, you need to do some dusting

Don’t mind me just among some of my favorite things in this world, including @avec

discovered an old 2TB hard drive earlier today and I've yet to feel this much of nostalgia b4.

Actual whimsical-maxxing on a nice Sunday in Toronto for running -> swimming -> reading Munkres' Topology -> climbing -> exercise free will and bike around toronto for 1 hour -> clean houses (forgot to take photos cuz i'm actually just living)

@simon_mo_ Wait can i grab one of the hats :))

Very importantly! We have a beautiful patio on Market Street.

I think you might be interested into epiphenomenal qualia, as in subjective/conscious experience that are causally inert. Arts are very much this, and there is a disconnect between our mathematical substrate versus what is innately ours. This also known within the neuroscience-sphere where we can construct every components of the brain, yet the "conscious" experiment is very much unknown and we have yet to figure out an answer. I understand that your argument surrounds the ideas of sampling in a large search space to find something that is truly unique. This is very much possible, similar to move 37 from alphago, but at the end of the day they are tools to accelerate findings, rather than finding-newer-thing-in-itself, which might not be a bad thing also.

my sister is one of the most talented artists i know. she self taught multiple mediums before she even started high school and has done everything from hand drawn oil pastel animations to recording her own music. and she’s vehemently anti gen AI. whenever i’ve probed her on why, she says art is inherently human and anything made by AI lacks the human touch to qualify. i can see where she’s coming from but i disagree with this sentiment. i think AI is merely another medium that is up to us to wield, whether for artistic purposes or others. it makes me sad to see so many creatives who refuse to even consider the possibility that AI could be used produce something with artistic merit - not just for the sake of their own creativity, but because their input is critical to how the landscape of creative AI will be developed. for a community built on challenging ideas, institutions, and conventions, i think it is critical for artists to be leading the conversation beyond just saying “don’t do it, it’s bad”. creative AI is being molded right in front of us and if artistic insiders don’t step in, the conversation will ultimately be driven by the default perspective of the institutions they feel undercut by in the first place. obviously i love my sister and want her to succeed in her artistic endeavours. it just makes me nervous to see her and other people like her not be willing to engage in something that is and will continue to shift the paradigm when the rare opportunity for them to steer the ship is right there.

Kindly request Codex the app to have the capabilities to references chat-across-projects-slash-sessions thank you gpt-5.5-pro go brr but I also want for it to work across project and reference chat

Rookie number

@badlogicgames Rtk is also very good

TIL about "caveman mode" to "save tokens". how many tokens in a session are actually model output? i think i'll become a gardener.

@badlogicgames Im waiting max studio m5

my m1 max is falling apart. should i wait or buy some beefed out m5? i'm so torn.

@jnnnthnn Clearly not real

every user sends us an email when they onboard i respond to every one of them sometimes i respond so fast that people need a proof of humanity

@deepfates I hail from LessWrong. Please assume I understand all of the vernacular type beat

we're long past the age of prompt engineering being yelling at the model what to do and not do. prompt engineering now knowing The exact vocabulary terms of every domain and how to use them in context. Like Catch Me If You Can

@profjoeyg @simon_mo_ congrats!

Today I’m excited to congratulate @simon_mo_ on an outstanding PhD thesis defense on his work exploring the design of Inference Serving Systems. 🎉 Simon has been working on inference systems with me for nearly a decade -- long before most people even considered inference serving a research problem worth studying. Over that time, he helped drive inference systems projects spanning Clipper, @raydistributed Serve, and now @vllm_project. Together, these systems helped define the modern inference serving stack that powers today’s AI applications. Beyond being an exceptional researcher, Simon has also been a remarkable team and community builder, especially through his leadership on vLLM and the open-source ecosystem around it. Along with my colleagues @istoica05 and @koushik77, I am excited to see Simon leading @inferact as CEO and helping shape the future of inference systems and AI infrastructure. Congratulations, Simon!

@badlogicgames @mitsuhiko a lot of lessons learnt along the way, but modularization of plugins (esp. in vLLM) is definitely a good way forward to keep the core very lean, as well as the recent introduction of Triton-first sampling for Model Runner makes it somewhat simplier to work with.

@aarnphm @mitsuhiko definitely not for the faint of heart, and also quite hard given the speed at which things are going.

today was the day @mitsuhiko had his first look into the source code of a bunch of inference engines (not going to name names) and like me a few months ago, he now has stared into the abyss. we will never be the same.