Replying to
@kcarruthers@infosec.exchange check the guide and the complementary resources about this topic that we are publishing under CC licenses at #tagoross
Http://www.tagoross.com/content-release
#opensourceai #aiopenness
#opensourceai
11 posts · Last used 1d
"... China now restricting access to certain open source models."
#MaximilianNegele, 2026
If all the points in the OSD (Open Source Definition) are being followed, access can't be restricted. So if it can be restricted, it's not Open Source, by definition.
At least, that was the case, until Stefano Maffulli, as director of OSI, led the charge on drafting a new definition for "Open Source AI", by watering down the OSD until some MOLE training fit into it. *sigh*
#OpenSourceAI #AI #OpenSource
Boosted by @welcome@friends.deko.cloud
Hi everyone! Updated #introduction after >3yrs 😱
It’s a me, +mala! 🇮🇹🇬🇧
Academic at heart, I love teaching. Been doing #OpenSourceAI at @MozillaAI@mastodon.social for the last 3yrs now. I believe ppl should own their AI stack.
Projects:
- 3564020356.org (>25yrs 😅)
- #PicoGopher
- BYOTA (Build Your Own Timeline Algorithm)
- current maintainer of https://github.com/mozilla-ai/llamafile🦙📁
Interests: #bouldering #gopher #SelfHosting #reversing #ML #solarpunk
Blog: https://aittalam.github.io gopher.3564020356.org
💙
AI Sovereignty discussions have shifted as we discuss a Stack and "layered sovereignty" following the NVIDIA 5-layered cake 🍰. Shift from frontier models to individuals building the infrastructure layer, open source, via GitHub. 🛠️
Read the full report here:
https://openuk.uk/report/aiopenness2026/
#OpenModels #OpenWeights #AIOpenness #OpenSourceAI #OpenSource #AIInfrastructure #AIStack #Llama #NVIDIA #OpenUK #OpenHQ
NVIDIA, Microsoft, and 27 founding members launch the Open Secure AI Alliance to boost cyber defense, spurred by the Hugging Face autonomous attack.
#OpenSecureAI #NVIDIA #Microsoft #Cybersecurity #OpenSourceAI
NVIDIA, Microsoft, and Meta lead 25 firms signing an open-weight AI letter urging US policymakers to avoid premature limits on open models.
#OpenWeightAI #OpenSourceAI #NVIDIA #Microsoft #AIpolicy
The Arguments Against Open Source AI Are Very Bad
https://tombedor.dev/arguments-against-open-source-ai-are-very-bad/
#HackerNews #Tech #OpenSourceAI
Moonshot AI’s Kimi K3 — now at 2.8 trillion parameters — is reshaping expectations for open models. Our latest article outlines the technical advances, market appetite, and implications for researchers and product teams. Read the full analysis: https://wix.to/zoFDROW#AI#Research#OpenSourceAI#TechLeadership
#MachineLearning
China AI export controls could restrict model weights, training data, and chip designs as Beijing starts guarding frontier AI as a national asset.
#ChinaAI #ExportControls #AIRegulation #Semiconductors #OpenSourceAI
Z wyboru linków zainteresował mnie artykuł o Eustelli, czyli europejskiej bezpiecznej (czyżby?) altenatywie dla Gemini i ChatGPT.
Mam kilka wątpliwości. Eustella korzysta z modeli,
cyt: "Gemma 4 (Google), Qwen 3.5/3.7 (Alibaba Group), gpt-oss-120b (OpenAI) i Mistral (Mistral AI) i Flux"
Czy korzystanie z tych modeli jest rzeczywiście bezpieczne? Czy modele typu open-weight na prawdę są godne zaufania? To nadal są modele zamknięte, dające jedynie możliwość uruchamiania ich lokalnie i nie dają prawdziwej przejrzystości sposobów swojego działania oraz zasobów, na których były trenowane.
Czy ktoś, kto ma szerszą wiedzę na ten temat, może odnieść się do moich wątpliwości?
Więcej na temat problemów z open-weight:
https://opensource.org/ai/open-weights
Art. o Eustelli:
https://www.tabletowo.pl/eustella-europejska-alternatywa-dla-chatgpt-i-gemini/
#ai #opensourceai #openweightaimodels
You've seen all posts