siavash
As a IT professional and developer
As a IT professional and developer, I was genuinely excited to test the Even Realities G2 glasses. I bought them hoping for a discreet, contextual AI assistant that could support real-time translation and professional workflows. Unfortunately, after two weeks of rigorous testing, I decided to return them. Here is my technical breakdown of why the G2 feels like a premium prototype that was rushed to market. The Good: Optics and Wearability To be fair, the hardware design team has done a commendable job. The glasses are remarkably light, socially acceptable, and look like normal premium eyewear. The waveguide display is crisp, and the charging case feels sleek and well designed. From a pure hardware perspective, the G2 proves that display-enabled smart glasses can be wearable. However, a smart device is only as good as the software ecosystem supporting it, and this is where the G2 falls short. The Technical Dealbreakers The Infinite “AI Server Error” Loop The core selling point of these glasses is their AI functionality. In practice, however, the backend architecture felt highly unstable. I was repeatedly hit with “AI Server Error” prompts. For a device that relies on cloud-native processing, unstable server infrastructure means the glasses can quickly become an expensive, non-functional piece of hardware. Extreme Latency in Real-Time Translation Living in Sweden, I heavily tested the real-time translation feature. The latency was too high for real-world conversations. By the time the system processed the audio, generated the translation, and projected it onto the lens, the conversation had already moved forward. Poor Acoustic Performance and Microphone Sensitivity The marketing suggests a discreet AI assistant. However, in my experience, the microphone array struggled to register low-volume speech or whispers. To get the AI to recognize commands, I often had to speak at a normal or loud volume, which defeats the purpose of a discreet wearable in professional or public settings. False Ecosystem Promises: The BYOA Illusion Even Realities markets concepts such as “Bring Your Own Agent” and connecting your own AI stack. However, when I contacted technical support, I was told there is currently no private API access or true developer mode. In practice, users appear to be locked into Even Realities’ own model-switching system rather than being able to connect their own AI infrastructure. The Customer Experience and Return Process The customer experience was also frustrating. While I was trying to resolve my return request and clarify my EU 14-day return rights, I received an automated marketing email promoting “Trust, Context, and Privacy.” The contrast between the marketing message and the support experience felt disconnected. If you request a return, be prepared for a long email chain. In my case, support initially pointed to temporary server outages and hardware design limitations rather than acknowledging the practical problems I experienced. After strong pushback and reminders about credit card protection, they eventually waived the restocking fee, but they still deducted €30 for return shipping. The Even Realities G2 is a classic case of hardware being far ahead of software. If you are looking for a reliable, production-ready AI companion, this is not it yet. The cloud infrastructure feels fragile, the AI features are sluggish, and the support process is more defensive than customer-focused.
Response from Even RealitiesJun 3, 2026
Hi there, Thank you for sharing your experience and honest feedback. We're sorry to hear the G2 didn't meet your expectations, and we take your points seriously. Your insights have been shared with our team as we continue working to improve our products and services. If you'd like to discuss anything further, please feel free to reach out to us at [email protected].


