Javier Jardón

@0xSero @MiniMax_AI @Zai_org @Alibaba_Qwen @Kimi_Moonshot @arcee_ai @OpenAI @googlegemma @nvidia @allen_ai :) I always surprised it's not mentioned more often (specially within localai enthusiastics); I have found almost by coincidence! I guess is not as good as the latest open weight models?

@jjardon @MiniMax_AI @Zai_org @Alibaba_Qwen @Kimi_Moonshot @arcee_ai @OpenAI @googlegemma @nvidia @allen_ai Yes I love allen

Let me local AI pill you: 1. It sucks compared to SOTA 2. It can’t code so well 3. It can be a good agent 4. It can be great at chat 5. It can be fine as a researcher 6. It can be a great automation engine 7. It can be tuned however you want 8. It teaches you how the sausage is made 9. It works on a plane, or in an outage 10. It costs your electric bill + hardware 11. It is better than the AI we gave up coding for a year or 2 ago. Local AI is self defence, it is a go kit, it is a rebalancing of power. It’s delusional to think it approaches or will ever approach SOTA, the scale of private labs blows anything you can get for less than 25k USD out the water. Local AI is a bet that prices won’t stay this low, that private corporations with closed source weights can’t be trusted to stay consistent. I am more than happy to rent a Ferrari for dirt cheap, but i should also have a beater Toyota if I can afford it. Local AI is the car I can depend on to be there tomorrow, something that’s mine.

@0xSero @MiniMax_AI @Zai_org @Alibaba_Qwen @Kimi_Moonshot @arcee_ai @OpenAI @googlegemma @nvidia What about @allen_ai ? Seems they tried to go even further and not only do open weights but a open the training material and processes. Seems even more aligned with the open source must win vision

Thank you to @MiniMax_AI @Zai_org @Alibaba_Qwen @Kimi_Moonshot @arcee_ai @OpenAI @googlegemma @nvidia for making it possible to run AI at home. I’m very grateful for your contributions to open source

Memory bandwidth for local AI hardware matters a lot more than most people think People keep comparing boxes like this: model size vs memory capacity That is only half the story The better mental model is: > capacity = what fits > bandwidth = how hard it can breathe > software stack = how much of that you actually cash out You are buying a memory subsystem and then negotiating with physics Here is the current local AI hardware ladder: > RTX PRO 6000 Blackwell > 96GB > 1792 GB/s > RTX 5090 > 32GB > 1792 GB/s > RTX 4090 > 24GB > 1008 GB/s Raw single-card bandwidth king stuff Now Apple > Mac Studio M3 Ultra > up to 512GB unified memory > 819 GB/s > Mac Studio M4 Max > up to 128GB > 546 GB/s > MacBook Pro M5 Max > up to 128GB > 460 to 614 GB/s > MacBook Pro M5 Pro > up to 64GB > 307 GB/s > Mac mini M4 Pro > up to 64GB > 273 GB/s > MacBook Air M5 > up to 32GB > 153 GB/s Apple is not winning raw bandwidth vs top NVIDIA Apple is winning the: > “I want one quiet box with a stupid amount of usable memory” argument And that is still a very real argument Now another interesting new category > DGX Spark > 128GB unified memory > 273 GB/s > GB10 class boxes like ASUS Ascent GX10 > 128GB unified memory > 273 GB/s These are not bandwidth monsters They are coherent-memory NVIDIA CUDA appliances That matters Because 128GB in one box changes what fits locally, even if it does not magically outrun a 5090 once the same model fits on both + CUDA Then there is the one category that actually made x86 interesting again for local AI: > Ryzen AI Max / Strix Halo > up to 128GB unified memory > 256 GB/s > up to 96GB assignable to GPU on Windows This is also where the Framework Desktop matters Not “just another mini PC” This is one of the first mainstream x86 boxes where local AI starts feeling like a serious hardware class instead of a laptop pretending very hard Then the trap people keep falling into: Most “AI PCs” are not in this tier They are down here: > Snapdragon X Elite > 135 GB/s > Intel Lunar Lake > 136 GB/s > Snapdragon X2 Elite > 152 to 228 GB/s depending on SKU > regular Ryzen AI 300 class way closer to thin-and-light territory than Strix Halo These are fine machines But the AI sticker does not create memory bandwidth Physics is still in charge which is rude but consistent AMD discrete cards > RX 7900 XTX > 24GB > 960 GB/s > Radeon PRO W7900 > 48GB > 864 GB/s > Radeon AI PRO R9700 > 32GB > 640 GB/s Not the CUDA default answer but definitely not irrelevant Intel is interesting now too > Arc Pro B65 > 32GB > 608 GB/s > Arc Pro B60 > 24GB > 456 GB/s And then there is Tenstorrent > Tenstorrent Wormhole n300 > 24GB > 576 GB/s > Tenstorrent Blackhole p150 > 32GB > 512 GB/s Not mainstream but absolutely relevant if you care about alternative and opensource local AI stacks So what does all of this actually mean? It means the local AI market is really five different markets wearing the same buzzword > fastest raw speed when it fits discrete NVIDIA > biggest one-box memory story Apple Ultra > coherent NVIDIA appliance DGX Spark / GB10 > first x86 unified-memory contender Strix Halo / Ryzen AI Max > oss stack Tenstorrent That is why people keep talking past each other A 5090 can absolutely embarrass a lot of unified-memory boxes if the model fits A Mac Studio M3 Ultra can fit things a 5090 cannot dream of fitting in one card A DGX Spark is interesting because it is compact coherent NVIDIA with 128GB & 273 GB/s + CUDA A Strix Halo box is interesting because it finally gives x86 a real answer to “what if I want big local models in one machine without going full workstation GPU?” Now Stop asking: > which box is best? Start asking: > what must fit? > what bandwidth tier do I need? > what software stack do I trust? > which bottleneck am I buying? That is how you stop guessing That is how you actually design a local AI system And yes most people still need to Buy a GPU

El orgullo de trabajar en una de las mejores centrales nucleares del mundo. Almaraz, que pretende cerrar el Gobierno entre 2027 y 2028 por razones políticas, genera el 7% de la energía eléctrica de España, constante y fiable, baja en emisiones de CO₂ y segura, según el CSN y organismos internacionales como WANO y el OIEA. Almaraz es una central tecnológicamente avanzada, gracias a inversiones anuales de más de 40 millones de euros, y está preparada para seguir operando, como North Anna (su central de referencia en EEUU), que tiene licencia para 80 años. Almaraz es un ejemplo mundial en seguridad, gracias a su equipo de 3000 personas (empleos directos e indirectos) altamente capacitadas, comprometidas y orgullosas, como muestra el vídeo.

@AnttiSaarnio I wonder: would jolla consider flatpak as a way to distribute apps? It will give immediate access to a lot Linux apps already available for mobile

MWC2026 was incredible. Over 2,000 people visited our stand, 100,000+ visited Jolla shop during the week, and we got overwhelming support from media, community and visitors who want an independent European mobile OS to exist. Thank you. Alongside all the love, I also see some hate from other alternative OS projects. I feel sad about this, as we don’t compete against other alternative OSes. We compete against the big tech and their dominance over the people. So let me tell you who we are and where we come from. Jolla was founded in 2011 to continue Nokia's MeeGo OS, which Microsoft killed after acquiring Nokia. It was a gigantic project for a startup. We quickly grew to 120 people, burning €1M per month with no major VC backing — Finland was not Silicon Valley back then. I personally became a fundraising machine: every 6 months I closed a round, and the morning after closing I started planning the next one, because I knew we'd be out of money in 6 months if I didn't. The community saved us multiple times. Our first phone pre-order campaign in 2013 was the most successful crowdfunding campaign of its time — it rescued the company after we lost our chipset vendor partnership overnight. But a €1M monthly burn with no major partnerships is unsustainable. We started gradually losing control to investors with different agendas. Not overnight — gradually. By 2017 I had lost control of the company and left. After Russia's invasion of Ukraine in 2022, I started negotiating to bring Jolla back into Finnish ownership. I spent 14 months in direct negotiations with Rostelecom — the Russian state operator that held the controlling stake. Eventually I had to accept we were going nowhere. May 2023, I walked away from the table. June 2023, Jolla entered voluntary restructuring. August 2023, the Municipal Court of Pirkanmaa ruled that Jolla's business, IP and personnel could be sold to a new company. I established that company — temporarily named Jollyboys Oy (I had to give it a name before we had rights to the Jolla brand, and I have sometimes weird humor). In November 2023, Jolla's creditors — the Finnish state, a major Finnish pension fund, and two leading Finnish law firms — voted in favour. Jolla was reborn as a 100% Finnish-owned company. Rostelecom appealed. They lost. So yes, Jolla has a history. We have always been open about it. None of it was hidden. Is Sailfish OS perfect? No. There is a lot to improve and a long way to go. We will gradually open source it, step by step. It may not yet be where the hardcore privacy community wants us to be — but we will get there. In the meantime, we can serve ordinary people who want privacy and real control over their data. A physical switch that cuts camera, microphone and Android apps with one press of a button is a great start. I have been in this business for 15 years. I have seen many OS projects come and go, or end up controlled by Google. The only way alternative operating systems win is if we help each other. Spreading hate helps no one. Jolla does not participate in spreading hate. We only spread love. The dogs bark, but the caravan moves on.

DC-ROMA RISC-V Mainboard III for @FrameworkPuter Laptop 13 — Early Access is live! 👉 store.deepcomputing.io/products/dc-ro… #SpacemiT #k3 #RVA23 support $99.99 to reserve #RISCV #OpenHardware

9,500 pre-orders. Thank you. ❤️ The pre-order price is still available — but only until 28 February 2026 at 23:55 (Helsinki time). That is less than 24 hours from now. If you have been waiting, now is the time. Secure your Jolla Phone at the pre-order price before the window closes. 🔗 jolla.com/phone #Jolla #SailfishOS #DigitalSovereignty #PrivacyFirst #European #CommunityPowered

9,000 pre-orders. Time to celebrate. 🎉 This is not just a number. It is 9,000 people who chose privacy, independence, and a different kind of phone. People who decided that their data is theirs. People who believe something better is possible. Secure your Jolla Phone at the special batch price before February ends. If you have been considering it, this is the moment. Let's make it together. ❤️ jolla.com/phone #Jolla #SailfishOS #DigitalSovereignty #European #PrivacyFirst #Linux #CommunityPowered

Here's a look at the five-year evolution of the Comet! Each iteration has been a vital lesson in understanding what a Linux handheld should be. If you're just joining us, we're entering the final week of our Kickstarter campaign with 2,700+ backers: kickstarter.com/projects/mecha…

What if one device held infinite possibilities? The Mecha Comet is modular, extensible and ready to adapt to whatever you can think up! Live on Kickstarter: kickstarter.com/projects/mecha…

The Comet is designed to be easy to open, fix and and re-assemble on your own :) We're releasing mechanical drawings and 3D CAD models, mainboard PCB schematics and reasonably priced spares. Longevity is not an afterthought for us.

This is a great report that provides a thoughtful, detailed and very well researched description of the risks of AI. It is essential reading for anyone who wants to write or talk about AI risks.

I just watched a really great conversation about the future of AI. Every politician should watch it before they join the lemmings saying that regulation of AI will interfere with innovation. youtube.com/watch?v=rGAA59…

5,000 Jolla Phones pre-ordered in one week, driven by the community. People showed there’s real demand for a European Linux phone built differently. The Jolla Phone reached 5,000 pre-orders organically, through community sharing and word of mouth. Thank you. We couldn’t have done this without you. Next stop: 10,000 pre-orders to unlock The Other Half and open source the modular hardware & software interfaces. 👉 Pre-order the Jolla Phone: jolla.com/phone 📰 Read the press release: jolla.com/content/upload… #Jolla #SailfishOS #DeGoogle #european #DigitalSovereignty #CommunityPowered

Wishlist Super Woden:Rally Edge NOW! 😃 store.steampowered.com/app/3218630/Su…

Con la ayuda de @ViJuDaGD, @Elvendwarf2 y Pablo Monroy de @OP_Team_ he intentado hacer una pequeña reseña del estado actual de la industria del videojuego en Galicia, y en concreto de una ciudad emprendedora como Vigo 👇🏽 atlantico.net/vigo/industria…

On my way to one of the best conferences of the year! Get all your devices with #GNOMEOS ready! :) See you all in Brescia! #GUADEC2025

Help make it happen for Liberux NEXX (linux phone) on @indiegogo igg.me/p/3287968/twtr…

Matthias and Florian explaining us in @LinuxAppSummit something really exciting; #gtk on #android !! (with demo included!) #LAS2025

Thank you @codethink for supporting #LAS2025! Learn more and find out how your company can become a LAS sponsor: linuxappsummit.org/sponsors/