Jolla<\/a> has taken the official wraps off the first version of its personal server-based AI assistant in the making. The reborn startup is building a privacy-focused AI device \u2014 aka the Jolla Mind2<\/sup><\/a>, which TechCrunch exclusively revealed at MWC back in February<\/a>.<\/p>\n At a livestreamed launch<\/a> event Monday, it also kicked off preorders, with the first units slated to ship later this year in Europe. Global preorders open in June, with plans to ship later this year or early next.<\/p>\n In the two+ months since we saw the first 3D-printed prototype of Jolla\u2019s AI-in-a-box, considerable hype has swirled around other consumer-focused AI devices, such as Humane\u2019s Ai Pin<\/a> and the Rabbit R1<\/a>. But early interest has deflated in the face of poor or unfinished user experiences \u2014 and a sense that nascent AI gadgets are heavy on experimentation, light on utility.<\/p>\n The European startup behind Jolla Mind2<\/sup> is keen for its AI device not to fall into this trap, per CEO and co-founder Antti Saarnio. That\u2019s why they\u2019re moving \u201ccarefully\u201d \u2014 trying to avoid the pitfall of overpromising and under delivering.<\/p>\n \u201cI\u2019m sure that this is one of the biggest disruptive moments for AI \u2014 integrating into our software. It\u2019s massive disruption. But the first approaches were rushed, basically, and that was the problem,\u201d he told TechCrunch. \u201cYou should introduce software which is actually working.\u201d<\/p>\n The feedback is tough, but fair in light of recent launches. <\/p>\n Saarnio says the team is planning to ship a first few hundreds of units (up to 500) of the device to early adopters in Europe this fall \u2014 likely tapping into the community of enthusiasts it built up around earlier products such as its Sailfish mobile OS.<\/p>\n Pricing for the Jolla Mind2<\/sup> will be \u20ac699 (including VAT) \u2014 so the hardware is considerably more expensive than the team had originally planned. But there\u2019s also more on-board RAM (16GB) and storage (1TB) than they first budgeted for. Less good: Users will have to shell out for a monthly subscription starting at \u20ac9.99. So this is another AI device that\u2019s not going to be cheap.<\/p>\n The Jolla Mind2<\/sup> houses a series of AI agents tuned for various productivity-focused use cases. They\u2019re designed to integrate with relevant third-party services (via APIs) so they can execute different functions for you \u2014 such as an email agent that can triage your inbox, and compose and send messages. Or a contacts agent which Jolla briefly demoed at MWC that can be a repository of intel about people you interact with to keep you on top of your professional network.<\/p>\n In a video call with TechCrunch ahead of Monday\u2019s official launch, Saarnio demoed the latest version of Jolla Mind2<\/sup> \u2014 showing off a few features we hadn\u2019t seen before, including the aforementioned email agent; a document preview and summarizing feature; an e-signing capability for documents; and something new it\u2019s calling \u201cknowledge bases\u201d (more below).<\/p>\n The productivity-focused features we saw being demoed were working, although there was some notable latency issues. An apologetic Saarnio said demo gremlins had struck earlier in the day, causing the last-minute performance issues.<\/p>\n Switching between agents was also manual in the demo of the chatbot interface but he said this would be automated through the AI\u2019s semantic understanding of user queries for the final product.<\/p>\n Planned AI agents include: a calendar agent; storage agent; task management; message agent (to integrate with third-party messaging apps); and a \u201ccoach agent\u201d, which they intend to tap into third-party activity\/health tracking apps and devices to let the user query their quantified health data on device.<\/p>\n The promise of private, on-device processing is the main selling point for the product. Jolla insists user queries and data remains securely on the hardware in their possession. Rather than \u2014 for example if you use OpenAI\u2019s ChatGPT \u2014 your personal info being sucked up into the cloud for commercial data mining and someone else\u2019s profit opportunity\u2026 <\/p>\n Privacy sounds great but clearly latency will need to be reduced to a minimum. That\u2019s doubly important, given the productivity and convenience \u2018prosumer\u2019 use-case Jolla is also shooting for, alongside its core strategic focus on firewalling your personal data.<\/p>\n The core pitch is that the device\u2019s on-board circa 3BN parameter AI model (which Saarnio refers to as a \u201csmall language model\u201d) can be hooked up to all sorts of third-party data sources. That makes the user\u2019s information available for further processing and extensible utility, without them having to worry about the safety or integrity of their info being compromised as they tap into the power of AI.<\/p>\n For queries where the Jolla Mind2<\/sup>\u2018s local AI model might not suffice, the system will provide users with the option of sending queries \u2018off world\u2019 \u2014 to third-party large language models (LLMs) \u2014 while making them aware that doing so means they\u2019re sending their data outside the safe and private space. Jolla is toying with some form of color-coding for messages to signify the level of data privacy that applies (e.g. blue for full on-device safety; red for yikes your data is exposed to a commercial AI so all privacy bets are off).<\/p>\n Saarnio confirmed performance will be front of mind for the team as they work on finessing the product. \u201cIt\u2019s basically the old rule that if you want to make a breakthrough it has to be five times better than the existing solutions,\u201d he said.<\/p>\n Security will also absolutely need to be a priority. The hardware will do things like set up a private VPN connection so the user\u2019s mobile device or computer can securely communicate with the device. Saarnio added that there will be an encrypted cloud-based back-up of user data that\u2019s stored on the box in case of hardware failure or loss.<\/p>\n Which zero knowledge encryption architecture they choose to ensure no external access to the data is possible will be an important consideration for privacy-conscious users. Those details are still being figured out.<\/p>\n One big criticism that\u2019s been leveled at early AI devices like Humane\u2019s Ai Pin and the Rabbit R1 takes the form of an awkward question: Couldn\u2019t this just be an app? Given, y\u2019know, everyone is already packing a smartphone.<\/p>\n It\u2019s not an attack line that obviously applies to the Jolla Mind2<\/sup>. For one thing the box housing the AI is intended to be static, not mobile. Kept somewhere safe at home or the office. So you won\u2019t be carrying two chunks of hardware around most of the time. Indeed, your mobile (or desktop computer) is the usual tool for interacting with Jolla Mind2<\/sup> \u2014 via a chatbot-style conversational interface.<\/p>\n The other big argument Saarnio makes to justify Jolla Mind2<\/sup> as a device is that trying to run a personal server-style approach to AI processing in the cloud would be hard \u2014 or really expensive \u2014 to scale.<\/p>\n \u201cI think it would become very difficult to scale cloud infrastructure if you would have to run local LLM for every user separately. It would have to have a cloud service running all the time. Because starting it again might take like, five minutes, so you can\u2019t really use it in that way,\u201d he argued. \u201cYou could have some kind of a solution which you download to your desktop, for example, but then you can use it with your smartphone. Also, if you want to have a multi-device environment, I think this kind of personal server is the only solution.\u201d<\/p>\n The aforementioned knowledge base is another type of AI agent feature that will let the user instruct the device to connect to curated repositories of information to further extend utility.<\/p>\n Saarnio demoed an example of a curated info dump about deforestation in Africa. Once a knowledge base has been ingested onto the device it\u2019s there for the user to query \u2014 extending the model\u2019s ability to support them in understanding more about a given topic. <\/p>\n \u201cThe user [could say] \u2018hey, I want to learn about African deforestation\u2019,\u201d he explained. \u201cThen the AI agent says we have one provider here [who has] created an external knowledge base about this. Would you like to connect to it? And then you can start chatting with this knowledge base. You can also ask it to make a summary or document\/report about it or so on.<\/p>\n \u201cThis is one of the big things we are thinking \u2014 that we need to have graded<\/em> information in the internet,\u201d he added. \u201cSo you could have a thought leader or a professor from some area like climate science create a knowledge base \u2014 upload all the relevant research papers \u2014 and then the user\u2026 could have some kind of trust that somebody has graded this information.\u201d<\/p>\n If Jolla can make this fly it could be pretty smart. LLMs tend to not only fabricate information but present concocted nonsense as if it\u2019s the absolute truth. So how can web users surfing an increasing AI-generated internet content landscape be sure what they\u2019re being exposed to is bona fide information?<\/p>\n The startup\u2019s answer to this fast-scaling knowledge crisis is to let users point their own on-device AI model at their preferred source\/s of truth. It\u2019s a pleasingly human-agency-centric fix to Big AI\u2019s truth problem. Small AI models plus smartly curated data sources could also offer a more environmentally friendly type of GenAI tool than Big AI is offering, with its energy draining, compute and data heavy approach.<\/p>\n Of course, Jolla will need useful knowledge bases to be compiled for this feature to work. It envisages these being curated \u2014 and rated \u2014 by users and the wider community it hopes will get behind its approach. Saarnio reckons it\u2019s not a big ask. Domain experts will easily be able to collate and share useful research repositories, he suggests.<\/p>\n Jolla Mind2<\/sup> spotlights another issue: How much tech users\u2019 experience of software is typically very far outside their control. User interfaces are routinely designed to be intentionally distracting\/attention-hogging or even outright manipulative. So another selling point for the product is about helping people reclaim their agency from all the dark patterns, sludge, notifications etc., etc. \u2014 whatever really annoys you about all the apps you have to use. You can ask the AI to cut through the noise on your behalf.<\/p>\n Saarnio says the AI model will be able to filter third-party content. For example, a user could ask to be shown only AI-related posts from their X feed, and not have to be exposed to anything else. This sums to an on-demand superpower to shape what you are and aren\u2019t ingesting digitally.<\/p>\n \u201cThe whole idea [is] to create a peaceful digital working environment,\u201d he added. <\/p>\n Saarnio knows better than most how tricky is it to convince people to buy novel devices, given Jolla\u2019s long backstory as an alternative smartphone maker. Unsurprisingly, then, the team is also plotting a B2B licensing play.<\/p>\n This is where the startup sees the biggest potential to scale uptake of their AI device, he says \u2014 positing they could have a path to selling \u201chundreds of thousands\u201d or even millions of devices via partners. Jolla community sales, he concedes, aren\u2019t likely to exceed a few tens of thousands at most, matching the limited scale of their dedicated, enthusiast fan-base.<\/p>\n The AI component of the product is being developed under another (new) business entity, called Venho AI. As well as being responsible for the software brains powering the Jolla Mind2<\/sup>, this company will act as a licensing supplier to other businesses wanting to offer their own brand versions of the personal-server-cum-AI-assistant concept.<\/p>\n Saarnio suggests telcos could be one potential target customer for licensing the AI model \u2014 given these infrastructure operators once again look set to miss out on the digital spoils as tech giants pivot to baking generative AI into their platforms.<\/p>\n But, first things first. Jolla\/Venho needs to ship a solid AI product.<\/p>\n \u201cWe must mature the software first, and test and build it with the community \u2014 and then, after the summer, we\u2019ll start discussing with distribution partners,\u201d he added. <\/p>\nAI agents living in a box<\/h2>\n
AI hardware with a purpose?<\/h2>\n
![\"\"](\"https:\/\/techcrunch.com\/wp-content\/uploads\/2024\/05\/Screenshot-2024-05-17-at-09.45.44.png?w=680\")