The First Update For Proton's Privacy-focused Chatbot Offers Major Performance Improvements

In July, Proton, nan institution down Proton Mail, released Lumo, a privacy-focused AI chatbot. Now, conscionable nether a period later, Proton has begun rolling retired Lumo 1.1 to some free and paying users, and according to nan company, nan updated adjunct "performs importantly amended crossed nan board" comparative to its predecessor.

Just really overmuch amended is nan caller type of Lumo? Proton claims it offers a 170 percent betterment successful discourse understanding, meaning nan chatbot is capable to much accurately reply questions based connected documents and information users stock pinch it. It's besides 40 percent amended astatine generating moving code, and 200 percent amended astatine moving done multi-step problems, arsenic good arsenic utilizing nan correct devices to tackle tricky prompts.

One area wherever nan original merchandise felt for illustration it wasn't up to nan level of nan title was erstwhile it came to providing answers astir existent events. Thanks to a amended web hunt tool, Proton says Lumo 1.1 should supply much meticulous answers while suffering from less hallucinations.

Proton doesn't opportunity really it measured those improvements, but nan institution did stock a fistful of screenshots that comparison nan responses generated by Lumo 1.1 and 1.0. If thing else, nan caller chatbot does a amended occupation of formatting its answers successful a measurement that's easy to follow.

As before, immoderate conversations you person pinch Lumo are encrypted, pinch nary logs shared pinch Proton aliases immoderate different company. Proton is besides releasing nan code for Lumo's iOS and Android apps truthful nan unfastened root organization tin verify nan company's information credentials.