October 6, 2026
TechCrunch reported that LibreOffice has confirmed it has no plans to bolt AI features onto its software by default, framing that restraint as a privacy-driven design choice rather than a gap in the roadmap. That's a notable thing for a software company to say out loud in 2026, when every vendor from your email client to your spreadsheet tool seems to be racing to bolt a chatbot onto the sidebar. LibreOffice is making the opposite bet: that a meaningful number of users are tired of AI they didn't ask for, running on data they didn't agree to send anywhere.
I think this is a smarter move than it looks at first glance. Most AI backlash coverage focuses on hallucinations or job displacement, but the quieter frustration -- the one LibreOffice is tapping into -- is about consent and control. Business buyers evaluating any tool right now should be asking not just 'does it have AI' but 'can I turn it off, and what happens to my data if I don't.' That's the same question we'd point teams toward when they look at ViibeStack's own security and trust center pages -- the AI feature is only worth it if you know exactly what it's doing with your information. LibreOffice won't win on raw capability, but 'no AI by default' is a real differentiator for regulated industries, cautious IT departments, and anyone who's had enough of AI theater bolted onto tools that worked fine before.
On the opposite end of the spectrum, TechCrunch reported that French lab Mistral AI has released Mistral Large 4, a new large multimodal model the company says is built to leapfrog both American and Chinese rivals. A trillion-parameter-class model is a genuine engineering flex, and Mistral has earned credibility as the one European lab consistently shipping frontier-scale work rather than just talking about sovereignty. For a business audience, though, the headline number matters less than what it signals about the competitive field: there are now at least three serious geographic poles -- US, China, and Europe -- all shipping frontier models within months of each other.
My honest take is that this is good news for buyers and bad news for anyone building a moat around a single foundation model. When Mistral, OpenAI, and the Chinese labs are all leapfrogging each other every quarter, the model itself stops being the differentiator -- what you build on top of it is. That's the argument we've made before about open vs. closed AI tradeoffs, and it's worth remembering as founders debate it on stage at TechCrunch Disrupt this month: the model layer is becoming a commodity faster than most roadmaps assume, so betting your whole business on exclusive access to one lab's weights is increasingly a risky strategy, not a safe one.
Sandwiched between those two bigger stories is a smaller one that's arguably more instructive for most businesses: TechCrunch reported that Pinterest's new AI-powered Beauty Guides translate hair and nail inspiration Pins into salon terminology, complete with estimated costs, appointment times, and maintenance needs. Nobody is going to call this a frontier breakthrough, and that's exactly the point. It's a narrow, well-scoped use of AI that solves a specific, annoying problem -- not knowing what to ask for or how much it'll cost -- without pretending to be a general-purpose assistant.
This is the pattern we keep seeing win in practice: AI that's wrapped around a specific workflow beats AI that's dropped into a product as a generic feature. Salons and service businesses that want this same logic -- translating customer requests into bookable, priced appointments automatically -- don't need to wait for a platform to build it for them; it's the kind of thing teams are already doing with workflow automation and tools built for marketing and booking. The lesson from Pinterest isn't 'add AI,' it's 'find the one annoying step in your customer's journey and automate just that.'
So which side of this are you on: are you more interested in tools that quietly do one job well, or in having access to the biggest, most capable model available -- and do you think most businesses can even tell the difference yet?
Sources