Tweets
-
Going full circle.🔄 Using LLM because natural language is an easy to use interface to do anything, including coercing to popular format (e.g. JSON), to basically query in natural language but have a non natural internal (here between agents, but not for the user) representation. https://x.com/MarroSamuele/status/1846924264764592252
(original)
-
RT @Oxfam: The super-rich are dining out on the planet-emitting more carbon in just 90 minutes than the average person does in a lifetime.…
(original)
-
Replying to @colin_fraser
Multi-dimensional goal-post moving.
(original)
-
Replying to @nhfoley
Couple of ASML High NA EUV and get going with your own Fab.
(original)
-
Replying to @drumr7 and @geerlingguy
FWIW https://x.com/utopiah/status/1453008004853403648 even though I do like to also have blog posts next to videos, if they don’t exist or transcripts aren’t great, this can help.
(original)
-
Replying to @utopiah
Last heard during https://www.santafe.edu/culture/podcasts/episode-3-what-kind-of-intelligence-is-an-llm warmly recommended
(original)
-
“we overestimate the impact of technology in the short-term and underestimate the effect in the long run.” Roy Amara https://x.com/Da_Real_Hugo/status/1853583057334235457
(original)
-
Replying to @N8Programs
cc @liquidizer, talk to paper?
(original)
-
Replying to @KenneyNL
2D map tiles
(original)
-
Replying to @utopiah
ToC :
Navigation and Network
TCP/TLS Handshake
HTTP Request/Response Cycle
Tokenization
DOM Tree Creation
CSSOM Tree Creation
Render Tree Creation
Layouting
Painting
Additional Reading
Summary
References
(original)
-
Don’t have time for @BrowserBook? Check “Exploring the browser rendering process” https://abhisaha.com/blog/exploring-browser-rendering-process by @abhisaha_1
(original)
-
Any interesting explorations, prototypes or writing, on a reactive filesystem?
Pushing inotify/fs watch() to their technical and conceptual limits?
(original)
-
Replying to @MVdWSoftware and @beatsigner
I sure do, would it also be best if I add you in CC to possibly facilitate?
(original)
-
Replying to @beatsigner
would happily come to the lab and do a short WebXR demo, IMHO quite aligned (and definitely inspired) by some of your work.
(original)
-
Replying to @utopiah
To clarify a bit the new stack is relying on https://git.benetou.fr/utopiah/text-code-xr-engine for XR interactions and a simplified (single node) and extended (e.g. file conversion) version of https://git.benetou.fr/utopiah/offline-octopus
(original)
-
Damn, that was 4 years ago😅 but… I managed to get back there with my own (local) stack : namely I can drag&drop a document in XR (here PDF again), get it to my device for “proper” reading🔖 “outside” of XR.
Feels like a set back but it opens a TON more possibilities📈🧩^🧩. https://x.com/utopiah/status/1250464627739279360
(original)
-
Replying to @chrisoffner3d and @thelinestudio
PS: apologies if I sound too critical, I do think it’s important hence why I’m pushing back a bit ;)
(original)
-
Replying to @chrisoffner3d and @thelinestudio
Can be about robotics or any other tech really.
(original)
-
Replying to @chrisoffner3d and @thelinestudio
So @chrisoffner3d maybe I missed a bit, and I’m not solarpunk expert, but can you recommend fiction (text, video, games, etc) that showcasing such empowerment, gradually being able to ignore, or even remove, everything that leads to enslavement through tech?
(original)
-
Replying to @chrisoffner3d and @thelinestudio
Most sci-fi is just that : “Hey, the good guys are coming to help, they have technology that will solve problems” only to discover gradually that they aren’t actually good & that the technology is just a mean to control.
Sadly most people “get” that in fiction, not while buying.
(original)
-
Replying to @chrisoffner3d and @thelinestudio
vision might not even be fully aware of why they reject it but I bet (and that’s just my surely biased) perspective that it comes from a fear, without even necessarily understanding the source, that it’s basically just yet another “scam” that will enslave them somehow.
(original)
-
Replying to @chrisoffner3d and @thelinestudio
Maybe I missed the part where they showcased autonomy. If they do, and they are not just fixed something trivial e.g. 3D printing a new casing for the robot but also changing say a CPU from their own foundry (!) them I’m plain wrong.
Otherwise I’d argue people who reject this…
(original)
-
On the danger of #solarpunk and romantic visions of emerging tech : https://x.com/utopiah/status/1853348481139720362
(original)
-
Replying to @chrisoffner3d and @thelinestudio
So… the vision should be hopeful but not when tools, regarding of the underlying technology, are built (e.g. with DRM), sold (e.g. with specific loans with repurposing) or maintained (e.g. preventing repairs) to alienate workers.
(original)
-
Replying to @chrisoffner3d and @thelinestudio
lose control of the mean of production. Same with seeds. It’s “high tech” and at the core of being able to deliver better yield… yet if they lose the next generation because Monsanto, by design, is preventing them to, they are gradually owning less and less.
(original)
-
Replying to @chrisoffner3d and @thelinestudio
any emerging technology in the process is not what matters. What matters is how the relationship evolves over time.
Buying John Deere equipment, even if very expensive, is a risk. If the equipment can not be fixed, the farmers aren’t actually owning them and thus gradually…
(original)
-
Replying to @chrisoffner3d and @thelinestudio
See my pinned tweet https://x.com/utopiah/status/1306566303839391744
The question isn’t “is progress good” or even “what is progress itself” but rather are inequalities growing, creating stronger dependencies, or can people genuinely enjoy the benefit of their work.
Have robotics or software or…
(original)
-
Replying to @chrisoffner3d and @thelinestudio
It’s beautiful, literally, but
it's not about robotics per se.⚠️Rather, it’s about who controls the means of production.
Is it the actual farmer and workers… or rather is the robot company that selling locked down tools that the farmer can’t fix independently?
(original)
-
LLM-driven search is amazing if you don’t care for veracity and provenance. https://x.com/emilymbender/status/1853279181012094999
(original)
-
RT @samerps: Very excited to share the release of my addon Blender FastHenry! Use Blender’s powerful modelling capability combined with ge…
(original)
-
Replying to @utopiah
Anyway, on the bigger economical picture, I do like the “Big Content” concept brought forward in e.g. https://x.com/utopiah/status/1646887667924103168
(original)
-
The worst part about this debacle is that DRMs hurt :
- content creators, making their content harder to be enjoyed
- consumers, unable to shape the content in a way they can actually enjoy it.
I’m not taking about anything “malicious”, rather basics like cropping margins.🤷♂️ https://x.com/utopiah/status/1849412004332646423
(original)
-
Replying to @JayHadHope
that’s just nice to have extras. The biggest “Wow…” moment, the actual mind f*ck 🤯, is putting the headset, leaning or even moving, and the virtual “world” is there, we are “in” it.
Everything else matters less. We already reached that stage years ago.
(original)
-
Replying to @JayHadHope
higher frame rate, entirely open ecosystem, open source stack, open hardware, and a ton more content… but that’s just “luxury”.
6DoF itself is VR!
Sure inside out tracking is damn convenient, hand tracking too. I wish for full-body tracking to arrive soon… but again…
(original)
-
Replying to @JayHadHope
I give workshops on programming and creativity using development in XR as a medium.
I use “old” Quest 1s, sometimes I show a Quest 3 only for color pass through.
Why? Precisely because IMHO it’s “good enough” indeed. I’d love to have a smaller form factor, higher resolution…
(original)
-
Replying to @utopiah
So I find this kind of language is detrimental to work that is on its own standing important BUT, maybe more importantly, on what it even means to be human. I believe as https://x.com/IrisVanRooij/status/1823447108571324553 suggests we need to be mindful of not degrading the very definition of humanness.
(original)
-
Replying to @utopiah
The research itself on multi-agent is fascinating, even important. FWIW I did tinker with e.g. ARGoS (which I used for https://x.com/utopiah/status/1571112471515254791 ) as “is a multi-physics robot simulator” but it is clearly about “robots”, not “digital human being”.
(original)
-
Anthropomorphisation at scale… what even is a “digital human being”? What is a “civilization” then?
Can I run 2 ls program, pipe them to each other at random interval and call them “agents”? If I add 100s or 1M or 1B of them does it become a “civilization”?🤷♂️ https://x.com/GuangyuRobert/status/1852397383939960926
(original)
-
Replying to @utopiah
PS: to prove my (limited still) understanding of the field of my 1st CV PoC dates back from 2011 https://fabien.benetou.fr/Portfolio/PostureMaintenance and for something a bit more visual and recent https://x.com/utopiah/status/1398177660107436033
(original)
-
Replying to @utopiah
the term AI has been engulfed in so much hype it often feels like everything and anything is labelled AI in order to grain traction.
Anyway, more to come! 😆
(original)
-
Replying to @utopiah
This isn’t a criticize of the great work showcased here. I have also made computer vision (CV) based interfaces and I do believe they are very interesting explorations.
What I’m trying to put forward here is the “AI first” kind of showcases. Sure CV is part of AI, yet sadly…
(original)
-
So… after my series
“AI is being used a glue (the most expensive and polluting ever!) when we lack API, files or protocols”
here starts the “AI is used a as a sexy interface for what we already know how to do”
Once it’s quite fun, even useful, but is AI a genuine enabler?🤷♂️ https://x.com/hamptonism/status/1852789095572890011
(original)
-
Makes a lot of sense indeed, and aligned with “convincing faster than good” characterization of the evolution of the field by David @Pierce, cf https://x.com/utopiah/status/1792253451344261170 namely optimization is NOT about factual truth, it is rather about sounding. https://x.com/GaryMarcus/status/1852508903084548547
(original)
-
Replying to @natalie_thenerd and @no1089
Some of my notes on both https://fabien.benetou.fr/Tools/Eink
(original)
-
Replying to @natalie_thenerd and @no1089
reMarkable Pro has one, also much thinner than PineNote (I have both) but also less open if you want to tinker.
Community on reMarkable is much more active though
(original)
-
Replying to @tovanbo
Dungeon Keeper in VR?
(original)
-
Replying to @utopiah
Literally starting from https://lichess.org/learn but letting build on everything, going much further and deeper with e.g. the API https://lichess.org/api and code https://github.com/lichess-org
(original)
-
Donated to @lichess. Absolutely flawless service, model Website for 2024 :
- it just works
- no ads
- no dark patterns
- all libre & open-source🐧
… and impeccable on-boarding, letting someone learn chess from basics, all interactive, without even needing an account.
♟️🫶
(original)
-
What grows a bubble isn’t air, it’s empty promises from people who benefit from the growth, regardless of outcomes. https://x.com/bcmerchant/status/1852402286204482013
(original)
-
The kind of Internet we need.🤩 https://x.com/itseieio/status/1852384111429562873
(original)