Having spent years as a passionate advocate of product minded and simplistic development, and particularly during my time championing Arc within our organization, I find myself reflecting deeply on Dia's emergence in the market. My experience with Arc was particularly memorable - I invested considerable time and political capital pushing it through our compliance channels, only to face an embarrassing setback when the project was discontinued. The young scale-up did not have the terms or policies our company would require and before that debate could even be finished - the product was gone.
Yet, to this day it is my personal driver. It still surpasses other browsers in terms of use and performance. This experience overall though has left me cautious about re-engaging with our compliance and IT teams until Dia can demonstrate substantial enterprise value.
After extensive testing of Dia, I find myself grappling with several fundamental concerns about its approach. The core selling point - a browser capable of cross-tab question answering - feels like a solution in search of a problem. While technically impressive, it doesn't address any critical pain points in our daily workflows (their demo videos may beg to differ but I would counter with “How often are you looking to synthesis information across 10 tabs?) . While our input points may be disparate, they are not neatly tied with browser tabs either. I would argue MCP has a far greater impact in handling that system to system communication. Moreover, the financial aspect cannot be ignored: between existing LLM subscriptions, AI-powered email clients, and various other tools, adding another subscription-based browser (or whoever pays the lunch) to our tech stack requires careful consideration of its ROI. At some point down the line, the bill comes due. Will all of these disparate AI apps stand the test, or will we enter another ecosphere where Netflix, HBO, Disney, etc all snatch you up at another cost.
Looking to improve your team’s estimates during cycles/iterations/sprints? Users of Linear will love Spades.
An application that’s easily brought in to help facilitate discussion and get an effort level set.
What's particularly striking about Dia is how it represents a departure from what made Arc truly special. Arc's genius lay in its subtle, thoughtful features that seamlessly integrated into our workflow - focusing on what you wanted to do. The ability to instantly copy links with a simple CMD + Shift + C, the intelligent handling of frequently accessed applications, and the native Pull Request tracking - these weren't flashy features, they fundamentally improved our daily interactions with the browser. Arc managed to enhance productivity without drawing attention to itself or underlying AI. It added in AI to areas where the usage of an LLM required no thought on the part of the user, creating what I can only describe as a "beautiful flow state" for users.
The loss of Arc's gesture-based summarization feature is particularly noteworthy. While Dia suggests using its chat interface as an alternative, this represents a regression in user experience, in my opinion. The immediate, intuitive nature of Arc's implementation has been replaced with a more cumbersome interaction model that breaks the natural flow of browsing, and requires a user to know exactly what they want and how to phrase that to the LLM. “Pinch to summarize” is accessible to far more than ‘craft a summarization prompt’. Dia, as I was led to believe, is meant to be more accessible to users than Arc was. The “just enter a prompt” approach goes wildly in the other direction. Each user is faced with an empty box asking them what they need done.
Perhaps most frustrating is Dia's approach to personalization. Rather than offering meaningful customization options that enhance productivity, users are confronted with an overwhelming array of choices that feel more like a quiz than a browser setup. Questions like "Who do you write like?" and "Who inspires you?" seem disconnected from the practical needs of a professional user. I read the Financial Times every morning, that doesn’t mean I need a monetary angle for every site; these prompts feel particularly out of place. The experience reminds me of standing in a pharmacy's headache medicine aisle, overwhelmed by choices without clear differentiation or purpose. The paradox of choice at the most inopportune moment. My headache only growing worse as I fail to discern between the 400 options which might be best for me. I don’t need my browser to do everything, hidden away behind the right keywords. I need it to guide me elegantly towards what I need in that exact moment.
No posts
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.