Jakub Zavrel
@jakubzavrel.bsky.social
Founder/CEO of Zeta Alpha (@zeta-alpha.bsky.social) / Founder @Textkernel / AI for a Deep Understanding of Work / Machine Learning / NLP / Semantic Search / Signal / Noise
Doesn’t it start to sound like they are selling the same thing multiple times and using it for xAI as well?
Reflection AI will pay $150 million a month beginning July 1, 2026 through 2029 for immediate access to Nvidia's latest GB300 AI chips and supporting hardware across SpaceX's Colossus 2 data center near Memphis, Tennessee.
Zeta Alpha's new Superagents: Multi-agent orchestration, skills and memory are here to deliver the future of complex knowledge work. Connected to your enterprise knowledge, customized to your expertise, and secure within your own IT Read the release blog here: www.zeta-alpha.com/post/zeta-al...
AI for Industrial Knowledge, 3D CAD and Sovereignty. What Zeta Alpha learned at Hannover Messe 2026. Read the full blog post here: www.zeta-alpha.com/post/ai-for-...
Ik vind deze campagne erg belangrijk - doe je ook mee? actie.degoedezaak.org/petitions/st...
Stop de Amerikaanse overname DigiD!
Onze privé-gegevens zijn bij overname door een Amerikaans bedrijf niet veilig! Dit lijkt ons niet de veilige manier nu president Trump in Amerika aan het roer staat. Dit kan misgaan! Als de Nederland...
actie.degoedezaak.org
New paper out in Proceedings of the Royal Society B: we apply linguistic tools to sperm whale vowels. The result: sperm whale vowels do not just look like human vowels. They also behave like them. We found several parallels. Like in Latin, whales have short and long vowels.
This Thursday at 16:00 we are back with Zeta Alpha's Trends in AI. Together with Edwin Rijgersberg (ex-NFI LLM expert, author of GEITje) we will be dissecting the latest AI news, models, papers, and the current state of Sovereign AI, Local LLMs and Open Source AI. luma.com/4pulz31n
Trends in AI by Zeta Alpha - Sovereign AI, Local LLMs and Open Source · Luma
Join us for the Zeta Alpha Trends in AI webinar - in person in Amsterdam (LAB42, Science Park) or online in San Francisco and everywhere else in the world.…
luma.com
Excited to connect w/ IR community at @ecir2026.eu in Delft this week! R&D on search, relevance & evaluation is the foundation any advanced AI deployment in knowledge-heavy domains. The future? Agentic search systems that plan, query, evaluate & iterate, built on proven IR fundamentals. #ecir2026
Going to @ecir2026.eu in Delft? Or need another reason to go there? Please drop by SCOLIA 2026, our Second International Workshop on Scholarly Information Access to enjoy some exciting presentations! Our programme is online now at sites.google.com/view/bir-ws/.... See you in Delft!
SCOLIA (BIR) - SCOLIA 2026
SCOLIA 2026, the Second International Workshop on Scholarly Information Access, will take place at ECIR 2026 in Delft, The Netherlands. The aim of the SCOLIA (SChOLarly Information Access) workshop, f...
sites.google.com
The plot thickens in European AI regulation with the recent adoption by European Parliament of the ‘Voss Report’. www.euronews.com/next/2026/03...
EU Parliament votes for ‘permanent’ fix for AI copyright
The report calls for a European register to list every copyrighted work used to train AI models and whether the artists have opted out.
euronews.com
Putting strategic oil reserves on the market when it might be the one time you will actually need them feels like selling your snow shovel just before a winter storm. www.marketwatch.com/story/oil-fu...
Why the IEA’s largest-ever release of oil reserves is pushing up crude prices
The largest-ever release of oil from the International Energy Agency’s emergency stockpiles lifted prices for the commodity up by nearly 5% on Wednesday.
marketwatch.com
🚨 De Belastingdienst gaat alles wat met BTW te maken heeft aan de Amerikanen overlaten en als ik zeg alles, dan bedoel ik alles! Servers, software, onderhoud en de VS krijgt ook toegang tot de andere servers van De Belastingdienst en zelf geen toegang. Lees hier 👉🏻 www.barbarakathmann.nl/btw-amerika/
Whoa! Who would've thought? ... The brain may do backprop after all. www.nature.com/articles/s41...
Vectorized instructive signals in cortical dendrites - Nature
Mice learning a neurofeedback brain–computer interface task show neuron-specific teaching signals in cortical dendrites, consistent with a vectorized solution for credit assignment in the brain.
nature.com
LLM Evals become the engine for self-improving systems that outperform hand-written prompts and multi-agent systems. @abcamara.com talks Friday at @irlab-amsterdam.bsky.social's SEA on how Deep Research agents evolve without human intervention. www.meetup.com/sea-search-e...
Diffusion LLM speed breakthrough by Stefano Ermon & Inception team: 1000+ tokens/sec on regular Nvidia GPUs at competitive reasoning output quality www.inceptionlabs.ai/blog/introdu...
Introducing Mercury 2 – Inception
Today, we're introducing Mercury 2 — the world's fastest reasoning language model, built to make production AI feel instant.
inceptionlabs.ai
It also takes a lot of energy to be a good human, but still you really should try.
"It also takes a lot of energy to train a human."
Canadian startup Taalas shows hardwiring Llama-8B weights into custom chip brings 50x inference speed vs GPU at 10x less energy consumed. Early stage, but very interesting. www.forbes.com/sites/karlfr...
Taalas Launches Hardcore Chip With ‘Insane’ AI Inference Performance
Taalas has launched an AI accelerator that puts the entire AI model into silicon, delivering 1-2 orders of magnitude greater performance. Seriously.
forbes.com
Important work as AI safety publications decline in hype cycle. This one went a bit under the radar in the press cycle
Today we’re releasing the International AI Safety Report 2026: the most comprehensive evidence-based assessment of AI capabilities, emerging risks, and safety measures to date. 🧵 (1/19)
Europe has a lot of talent! Just look at the Olympics.
It takes local champions that act to build digital sovereignty. That’s why Zeta Alpha is teaming up with OEDIV, part of the Dr. Oetker Group, to deliver Sovereign AI and to securely connect companies’ internal knowledge with locally run AI models. 🧵 1/6 Reach out via lnkd.in/dktJM9W to learn more.
Claude: For each example I can do a web search and then make a LLM call with the results... Me: Why an LLM call? Can't you just figure it out yourself? Claude: You're right, I am the LLM!
The pace of AI progress in China is extremely fast. Alibaba just dropped Qwen3.5. An open source VLM model competitive with the latest proprietary models. Superb efficiency and GPU footprint, 200+ languages, top reasoning, coding & multimodal tasks. qwen.ai/blog?id=qwen...
Two big breakthroughs this week in AI models: MiniMax M2.5 and Gemini 3 Deep Think.
The 12 Trillion dollar Molotov-Ribbentrop deal. Is this the breaking point? kyivindependent.com/russia-confi...
Russia confirms $12 trillion pitch to Trump tied to Ukraine deal — White House stays silent on proposals
Russia appeared to confirm on Feb. 13 the existence of a sweeping U.S.-Russia economic proposal known in Kyiv as the "Dmitriev package," days after President Volodymyr Zelensky first disclosed it publ...
kyivindependent.com
China keeps betting big on open source AI. State of the art performance with MIT license: GLM-5 release from Z.ai
Z.ai - Free AI Chatbot & Agent powered by GLM-5 & GLM-4.7
Meet Z.ai, your free AI-powered assistant. Build websites, create slides, analyze data, and get instant answers. Fast, smart, and reliable, powered by GLM-5.
z.ai