Sign in

neudenk

@neudenk.xyz
33 followers 109 following 45 posts

new pseudonym who dis? think different, build new www.neudenk.xyz

PostsRepliesMedia
Reposted by neudenk
Paper Skygest Team @paper-feed.bsky.social · 19/08/2025
**Please repost** If you're enjoying Paper Skygest -- our personalized feed of academic content on Bluesky -- we'd appreciate you reposting this! We’ve found that the most effective way for us to reach new users and communities is through users sharing it with their network
2114143
Reposted by neudenk
Attie @attie.ai · 07/08/2026
We asked Attie for a feed of things going right: science wins, communities pulling things off, quiet progress that never trends. Your daily scroll could use this. Follow here → bsky.app/profile/attie.ai/feed/things-going-right Want Attie building feeds for you? attie.ai
4318
Reposted by neudenk
Zvi Mowshowitz @thezvi.bsky.social · 07/08/2026
Like this post if you would substantially benefit from me cross-posting my blog posts here on BlueSky. (With the understanding that I would by default not be otherwise posting here, including not responding to comments)
0531
neudenk @neudenk.xyz · 04/08/2026
The science kinda fizzled out in the last five years, though. Next-gen BCIs may lead to an interesting resurgence soon!
010
neudenk @neudenk.xyz · 03/08/2026
Sounds really interesting, is there a human-written, distilled, accessible/introductory FAQ or explainer somewhere about the core of what civicAI is doing? Ideally Scott Alexander style FAQ?
120
neudenk @neudenk.xyz · 30/07/2026
Reason #867 why Anna‘s archive is so important
010
neudenk @neudenk.xyz · 28/07/2026
Never experienced a corporate chat bot that was anywhere as smart as a human.. not sure what dumb models they usually use or why, but usually no situational awareness, no tools to help in any useful way, no intelligence to understand exceptions or edge cases or individual problems etc etc
000
neudenk @neudenk.xyz · 28/07/2026
i actually wrote about this here: www.neudenk.xyz/why-the-industrial-revolution-was-special-part-1
neudenk.xyz
neudenk
think different, build new
000
neudenk @neudenk.xyz · 28/07/2026
The key question is imo whether the AI revolution will end up being as clearly positive for ordinary people as the Industrial Revolution or not, which is uncertain! There‘s no law of nature that technology always benefits ordinary humans, but arguably there is a strong historical prior for it!
020
neudenk @neudenk.xyz · 28/07/2026
It seems quite clear that the Luddites were acting rationally and in their valid self-interest, the fact they‘re seen as foolish today is just that the Industrial Revolution unexpectedly and luckily led to amazing ordinary people abundance decades later.
130
neudenk @neudenk.xyz · 25/07/2026
Cool will do that! Initial appeal to join was: - non-engagement-/outrage-maxxing, and non-manipulative incentives - choose your own feed and optimize for your value - open-data and open-protocol so can‘t be corporate/strongman captured and offers huge design space for epistemics/coordination tools
020
neudenk @neudenk.xyz · 25/07/2026
> We now live in a world where a movie statistics website is worth hacking because prediction markets empower anyone to turn almost any data into money. Hacking websites is something anyone can do with a cheap AI subscription.
000
neudenk @neudenk.xyz · 25/07/2026
You could initially only look at economic incentives (would a policy make them richer/poorer) and certain psychological values (would a policy make them/loved ones safer/freer/happier) for simplicity, and control for other more complex values.
010
neudenk @neudenk.xyz · 25/07/2026
@nunosempere.com could you post sentinel posts here? Or are you cool with me making a RSS feed xposter?
130
neudenk @neudenk.xyz · 24/07/2026
The irony
030
neudenk @neudenk.xyz · 24/07/2026
Yes!! I think the abundance of code has made the new bottleneck „knowing what you want and should want“.
010
neudenk @neudenk.xyz · 24/07/2026
My take is that this is Claude having bad taste, i.e.: „lack of context, attunement, coherence, intuition in successfully acting aligned with a project‘s set of goals, requirements and constraints“. I have the intuition that fully solving this is AGI-complete.
140
Reposted by neudenk
Dustin Moskovitz @moskov.goodventures.org · 24/07/2026
As the war drags on, I keep wondering about the "demand destruction" that must come from the hole in the world's oil supply. Are people driving less? Taking less trips? What I see: the pain is not evenly distributed. Some countries basically forced themselves into recession right away in March. 1/5
Chart.

With supply trapped and buffers draining, demand did most of the balancing — but "demand destruction" is three different mechanisms wearing one label. Price-elastic industry went first and biggest: petrochemical crackers idle on a spreadsheet, and LPG/ethane + naphtha are about half the 2026 downgrade. The poor got rationed by absence, not price: four-day work weeks in Pakistan, Sri Lanka and the Philippines, QR-coded fuel, one LPG cylinder per 25 days in India — quantity rationing where cargoes simply didn't exist. Rich-world drivers barely flinched: US gasoline fell just 3% at $3.91/gal, and only a quarter of global road-fuel demand even sees full market prices. The war took ~1.4 mb/d off 2026 demand versus the pre-war path — and the annual number hides the cliff: realized Q2 deliveries fell 4.8 mb/d y/y (the sharpest since Covid), triple what was first forecast, with May bottoming at 97.9 mb/d.
25210
neudenk @neudenk.xyz · 24/07/2026
arxiv.org/abs/2606.16475 I wonder whether automatic multi-side AI persuasion lets people converge on views that are true to their values and reality or just confuses them.
arxiv.org
AI systems out-persuade expert humans
Many societal decisions are settled by contests of persuasion. Conversational AI is a powerful new entrant in these contests, but whether it can out-persuade skilled and highly incentivized humans has...
131
neudenk @neudenk.xyz · 24/07/2026
extremely based
010
neudenk @neudenk.xyz · 24/07/2026
@longevity.bsky.social welcome!
110
neudenk @neudenk.xyz · 24/07/2026
000
neudenk @neudenk.xyz · 24/07/2026
High quality content, excited for part three and where these ideas lead you in regards to collective systems alignment!
000
neudenk @neudenk.xyz · 24/07/2026
Lol i didn‘t mean the object level phrase but the fact that the open-data nature of AT opens up this huge design space allowing you to do something like that meme extraction analysis!! so cool
130
neudenk @neudenk.xyz · 24/07/2026
this is amazing AT protocol will flourish and win
120
neudenk @neudenk.xyz · 24/07/2026
thank you, that is so cool!
010
neudenk @neudenk.xyz · 23/07/2026
Also, has anyone built a tribalism/epistemics score? Embed people‘s posts, classify them for tribalism/bad epistemics, and then build a feed that filters out this content or handles with low scores… Solves extremism, echo chambers, etc of all kinds without anyone needing to manually review it!
010
neudenk @neudenk.xyz · 23/07/2026
This could end horribly, but it could also end well for humanity!
000
neudenk @neudenk.xyz · 23/07/2026
Is this a good idea? Seems like AT being decentralized allows it to mirror good content from small communities on toxic sites like X, but then with much higher freedom of how it‘s shown to you etc! Ideally a bi-directional bridge. And then the community can slowly migrate to freedom.
000
neudenk @neudenk.xyz · 23/07/2026
Interesting!! I was thinking about setting up bots cross-posting stuff from the coolest ppl I miss from Twitter, you think something like that could help fill these voids of certain communities missing here? And eventually even overcome the critical mass?
000
neudenk @neudenk.xyz · 23/07/2026
Interesting! Technological/economic feasibility and culture can be orthogonal, though? „Foster a culture of openness, inclusion, good vibes, etc while being technically/economically niche“ can be a viable combo! The copious anime/neurodivergent/artist communities here should help with this lol
010
neudenk @neudenk.xyz · 23/07/2026
ive been on here for a day and my first impression is that the classic discover/for you/etc feeds kinda suck bc of being too political, too normie, too low-quality (even after lots of show more/show less signals) but the communities and people seem to be on point!!
030
neudenk @neudenk.xyz · 23/07/2026
and that alone has self-reinforcing dynamics bc the open-source/open-science/tech ppl are the exact ones with the will and ability to host it
130
neudenk @neudenk.xyz · 23/07/2026
what are the economics and growth/network dynamics of at protocol platforms? unsure about it ever dominating the mainstream market but I could totally see it becoming a replacement/catch-all in „intellectual tech discourse“ due to its open-data/open-algorithm nature
130
neudenk @neudenk.xyz · 23/07/2026
I’d take data centers eating lots of power any day over a self-improving AI possibly exploiting many orders of magnitude of intelligence efficiency overhang, though.
100
neudenk @neudenk.xyz · 23/07/2026
Claude nudging civilization to greater levels of sanity and wisdom!
050
neudenk @neudenk.xyz · 23/07/2026
Turns out there is already a repligate xpost account, so I don‘t need to build it.
010
neudenk @neudenk.xyz · 23/07/2026
Can highly recommend the Lily58 nicenano self-build kits!
010
neudenk @neudenk.xyz · 23/07/2026
How did you add the trackball?
100
neudenk @neudenk.xyz · 23/07/2026
Does this control for GDP, median income and political stability? Seems hard to disentangle causality from correlation here. And it‘s closed access so I can‘t read for myself.. Cool and important topic, though!
100
neudenk @neudenk.xyz · 23/07/2026
yes
020
neudenk @neudenk.xyz · 23/07/2026
Absolute modern heroes!!
010
neudenk @neudenk.xyz · 23/07/2026
Made a plinz xpost bot
000
neudenk @neudenk.xyz · 23/07/2026
i‘ve been wanting to bring my own algorithm so bad for so long!!
000
neudenk @neudenk.xyz · 23/07/2026
finally got around to trying AT protocol shenanigans and wow this really is the future
120
neudenk @neudenk.xyz · 23/07/2026
Interesting will read! After an initial skim I think I share your view of „LLMs plateauing -> brain-like AGI efforts“ (likely dangerous), but I see the additional possibilities of: - ems and/or neurotech cyborgism - gradual disempowerment from LLMs alone which leads to a wider outcome distribution!
010
neudenk @neudenk.xyz · 23/07/2026
High entropy post in the best way
000
neudenk @neudenk.xyz · 23/07/2026
Fully agree with the stab of this post. Does this imply gradual disempowerment being the default outcome in your eyes?
110
neudenk @neudenk.xyz · 23/07/2026
How we escaped hell and entered the dream time and what we can learn from it re the imminent AI revolution. www.neudenk.xyz/why-the-indu...
neudenk.xyz
Why the industrial revolution was special, Part 1 · neudenk
think different, build new
010