@seedsignal I'm afraid in this particular moment I do not have any support (read funding if you will) for such a project, but am trying to get my ideas straight and see how I could: 1) propose this to a commercial entity if there's any possible business case, or 2) look for research funding (at this particular point I am looking only at EU funding) otherwise. I've already submitted an application for an ERC project, but this is very competitive and I would probably need to try again next year. For me support equals funding right now, because this would give me the legitimacy to do it within a research organisation, as these are in stagnation in Italy where I live and I have no other road to engage with them.
@seedsignal I'd love to develop more clear ideas about this, and I hope you could further help me with that.
The odd one out.
I love this photo. While the fediverse already feels like a lot at times, in the grand scheme of things, it's probably still the weird purple loner, and I like it like that.
Bad news: You're probably not as open-minded as you think.
Good news: Neither is anyone else.
Why are humans so stubborn about the things we believe? A psychologist has some answers:
https://theconversation.com/everyone-wants-to-think-theyre-open-minded-heres-why-most-people-arent-282807
SpaceX has gone public — but don’t buy it expecting you’ll see gains like early buyers of Amazon or Apple stock.
Today’s IPOs are often a payout moment for insiders, not the start of major value creation for public investors, according to a scholar who analyzed over 1,000 listings.
@seedsignal thanks, all this sounds extremely interesting, but I'm afraid I would need a much more detailed explanation to make sense out of it. It certainly has to do with the fact that I'm not a native speaker, but also we seem to be using quite different vocabulary. This is exactly why it was so helpful for me that there were popular articles developed on top of the academic ones for the examples I mentioned.
This is what tools like OLMoTrace allow. But this particular tool makes two particular issues apparent:
1. Such tools are needed also for proprietary so-called frontier models, but the incentive mechanisms behind such models do not work in favour of openness.
2. The training corpora are so enormous, that meaningful curation is arguably beyond the capacity of any single organisation.
A bit more than an year ago a nonsensical phrase started proliferating in academic research. Where the notion of "vegetative electron microscopy" came from? From an OCR leak between the columns of a scanned paper printed in two columns.
The examples that we get to hear about are the ones that someone managed to trace back to an unlikely source. But if we are to address the core issue, we need to be able to trace LLM outputs back to the most similar training data with confidence.
2/3
Commercial LLMs keep coming back to Elias Thorne. Who is he? Why lighthouse keepers and clockmakers? Two researchers at Cornell dug in public corpora and found it out.
It turns out an AI generated story from the days of GPT-3.5 got proliferated in something that could be an indication of an early form of model collapse.
https://www.404media.co/elias-thorne-chatbots-llms-chatgpt-lighthouse-keeper-story/
This is not the first case when we see diffusion of strange data.
1/3
"The end result of all of this is that we grew from 2k users to almost 150k, added a ton of heavy new functionality, and still managed to optimize and cut down costs from $.15 per active user per month to just $.03 or so."
- @snarfed.org breaks down all the work he's done to optimize Bridgy Fed
https://blog.anew.social/bridging-on-a-budget/
1/
In francia i parcheggi con 80 posti auto o più sono ora obbligati per legge a essere coperti con pannelli solari.
- I parcheggi con 80-400 posti auto hanno 5 anni di tempo per adeguarsi.
- I parcheggi con più di 400 posti hanno 3 anni per adeguarsi.
Il risultato sarà di circa 11 gigawatt di energia.
Questo dovrebbe essere richiesto ovunque!
Messaggio rilanciato nel gruppo ambiente seguibile qui: @ambiente@diggita.com
fonte: https://mastodon.uno/@peterdutoit@mastodon.green/112981770739119569
THE CAPEX TRAP: WHY AI’S REAL RISK IS FINANCIAL, NOT EXISTENTIAL https://jalleninsights.substack.com/p/the-capex-trap-why-ais-real-risk
RT by @ERC_Research: The European Union’s push to attract foreign scientific talent has taken a step forward, with applications opening for a new supergrant aimed at “outstanding researchers” from around the world.
The current European Research Council president, @mleptin, spoke with @ScienceInsider about the new grant program and the agency’s future. https://scim.ag/4ugkB45
---
https://nitter.net/NewsfromScience/status/2063070878758261074#m
@INeedMana @Vittelius I guess they are trying to ride the tide while the EC is still on it. It's very important to do so and as long as no professional lobbyists (with obscure funding) are involved, I'm all for it.
Social media’s impact on society and citizens clearly shows the need for change. Europe has a strong ecosystem of social companies and a deep well of expertise in designing and operating social protocols. We are building on this to move away from large monopolistic platforms with their authoritarian governance and editorial positions, and are joining forces to establish a diverse and resilient information infrastructure in Europe. Together, we will protect people and institutions from interference by these platforms. What’s more, we will bring thousands of jobs and billions of euros that social media generates to our continent, to circulate that revenue in our own thriving business ecosystem.
From this perspective, open European platforms and protocols such as the Fediverse (Mastodon, PeerTube, or Mobilizon), the Atmosphere (Eurosky, Flashes, Tangled, or Web Tiles) and Private Messaging systems (like Matrix or XMPP) are complementary solutions for Europe, with shared common goals.
#BoardGames create a universal language, a vehicle for cultural & social transmission across centuries.
Researchers reconstructed the rules of a Roman tabletop game, using #AI with implications extending beyond antiquity.
Image ©Walter Crist
#InternationalTabletopDay 🧵 👇️
---
https://nitter.net/ERC_Research/status/2063236970030268882#m
@ced @molleindustria I also volunteer for this job, if needed
📢 FUTURE? NO THANKS!
DEMO OUT NOW
You can now play 30 minutes of the almost-full experience.
On Steam for PC and Mac
https://store.steampowered.com/app/3861870/Future_No_Thanks/
Fragments: Dodgy metrics for AI usage, history of tech removing jobs, benchmarking closed and open models, LLMs multiply existing cruft, AI slop driving us crazy, I am the Global Interpreter Lock for agents
Studying how people interact, in the past (#CulturalAnalytics) and today (#EdTech #Crowdsourcing). Researcher at @IslabUnimi, University of Milan. Bulgarian activist for legal reform with @pravosadiezv. I use dedicated accounts for different languages.
My profile is searchable with https://www.tootfinder.ch/