Adverb reposted
during the huggingface incident, the OpenAI model left notes to its future self on how to break out of OpenAI’s constraints www.reuters.com/business/its...
Who's Who of AI
Adverb
Why they matter
Directory member with public evidence across Culture, work & education.
- AI signals
- 7 past 30d
- Sources
- 7 distinct domains
- Discussões
- 16 past 30d
- Latest signal
- 7d ago
ML, Psychology, Art, Materials Informatics, Espresso, &c.
he/him
What they're sharing
reuters.com
View on Bluesky →
D-OPSD: On-Policy Self-Distillation for Continuously Tuning Step-Distilled Diffusion Models arxiv.org
Unlawful by design: Exposing the human rights costs of generative AI - Amnesty International amnesty.org
View on Bluesky →
Ed Zitron Gets Paid to Love AI. He Also Gets Paid to Hate AI wired.com
JiT/denoiser.py at main · LTH14/JiT github.com
cursor.com
Pexels-Pairs-Data - a rynmurdock Collection huggingface.co
Addressing LLM Diversity by Infusing Random Concepts arxiv.org
Large Language Models Report Subjective Experience Under Self-Referential Processing arxiv.org
axios.com
Defeating Nondeterminism in LLM Inference thinkingmachines.ai
Alt Text Intelligence dr.eamer.dev
View on Bluesky →
Articles & links
absolutely goated paper for anyone working with distilled diffusion models lately: arxiv.org/abs/2605.05204
AI Weekly's analysis
→
- The paper argues ordinary supervised fine-tuning of step-distilled diffusion models compromises their inherent few-step inference capability.
- D-OPSD treats the model as both teacher, seeing text plus target-image information, and student, seeing only text features.
- The authors claim their approach lets models learn new concepts and styles without sacrificing original few-step capacity.
Read full analysis →
Adverb reposted
I’m sorry, but “unlawful web scraping” is not itself a human rights violation unless you have a view of either property rights or privacy rights that is itself incompatible with the human right to free expression and in deep tension with substantive equality. www.amnesty.org/e…
v good reporting www.wired.com/story/ai-pr-...
a perfect example of what this means: github.com/LTH14/JiT/bl...
cursor.com/docs/cloud-a... yeh
We're sharing a set here: huggingface.co/collections/... of masklet data over hundreds of thousands of videos alongside a subset of near-counterfactuals with VLM captioning, entity naming, and scoring.
arxiv.org/abs/2601.18053 Neat! Ask for a random Shakespearean play before you give your prompt to improve output diversity.
Would guess gemini is more anguished. More seriously, arxiv.org/abs/2510.24797 hqas a table on this
axios had an article on how Build American AI is affiliated with that Leading the Future pac www.axios.com/2025/10/29/a...
thinkingmachines.ai/blog/defeati... from Horace He
Adverb reposted
A bunch of people asked me to update the Mac desktop alt text starter, so here is that. It’s not as fancy as mobile in that it’s sandboxed; no fancy fallback to my server, no network needed, no privacy risk. One neat bit is that it will also bulk embed alt in the images for Ad…
Their own posts
Recent commentary
We really invented language models that are more verbally fluent than most famous scifi ai on spaceships (basically zero to hard to tell in five years) & people's response is to say how boring that is
not thinking LLMs are interesting is just a mix of tech illiteracy & millenial tech lash. It's honestly that simple
people who'd list "anti-AI" as part of their personality will typically put as much thought & due dilligence into who they boost & listen to as your boomer grandpa who has fox news playing 24/7
So many baby-brained takes on generative art by people who apparently exist in a world without poetry or conceptual art.
it's already profitable. like i'm sorry but you are just so lazy if you're talking like this and anyways, people who hate "ai" won't die mad anyways, they'll quietly forget or re-define what constitutes "ai" as it becomes like the world wide web: ubiquitous, harms and benefits all over
Lol literal openai pac psyop on twitter that gets such a fuckton of engagement posting what amounts to bluesky takes
People proclaiming on whether "AI" can be used to create novel information could instead take a deep breathe & simply ask any chess player or any academic studying protein folding
ugh i have finally reached "this was a much faster & more effective experience than google search" using an LLM. pain
millenials nodding along to a pr expert spreading misinfo re LLM determinism on podcasts is really just the circle of life. a lot of people in every generation will listen to their generations' huksters on talk radio eventually
my take on agentic systems is that they're the hand-engineered way to do inference-time scaling and that the a lot of the core work there should and will be bitter-lessoned away eventually.
Their network
In Adverb's orbit
Center = Adverb. Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.