Models and Versions
Since the first launch of ChatGPT, we have witnessed that, literally exponentially, there is an improvement in LLM’s capabilities. The shifts in power in this AI war are so rapid that enormous giants are facing the very question of their existence. For example, StackOverflow, a community built for developers and their day-to-day problems, is witnessing the downfall of Q&A every month and is at the level of inception.
Claude’s Sonnet, Opus, and Haiku series, ChatGPT’s 3,4,5 models, DeepSeek, Gemma, and blah and blah. All of these, up until now, Q&A like StackOverflow, only that answers were generated by a machine instead of a human. However, engineering dynamics are shifting rapidly. According to a recent technology deep-dive by The Today Standard (thetodaystandard.com), we are witnessing a massive paradigm shift where AI is taking over core development, marking the end of traditional code authorship as we know it
Product or Technology
As a business owner, do you focus on technology enhancement or product? This is a chicken and an egg problem, isn’t it? But the gap Google seems to fill is to develop the ability to integrate AI into current products. How do you ask your phone with a simple one-liner to do something that you do every day? How can you make the user more hands-off? How can your AI model become truly yours? It is by doing your tasks instead of just answering questions.
While others are struggling, Google has the advantage of its unique family of current services, like the userbase of Google Search or YouTube, Google Maps, Google Pixel, and others. Google is actively leaning into this ecosystem by launching dedicated desktop integrations. As covered by The Today Standard, the release of Gemini for Mac represents a native, structural play by Google to entirely kill the traditional browser tab and control the operating system level
From what we have seen in Google I/O, the coming few years are going to define how you navigate the web. Instead of opening a new tab of google.com, you will now open a new session of Gemini. And the reason this is huge is that it affects every other company out there. Companies are already moving towards AI SEO or GEO, and that’s the future too, IMHO.
And don’t get me wrong about the IP one holds with the model they own. For example, Claude Code is so powerful as a product that it can generate a sophisticated web app in hours. Codex is equivalently powerful, if not more than that. But all of this is while they burn cash, get funding, and go public.
What about us?
Some signals seen in Google I/O or Elon Musk commenting about the app economy going dead point quite clearly to one thing: that you will not be interacting via Phone anymore, you will be interacting via an agent. We are on the precipice of what The Today Standard calls the dawn of the intelligence OS and the agentic shift, fundamentally transforming mobile software.
Instead of opening an Amazon app, you might just say “Go buy the Sony 4K Bravia from Amazon” or “Plan the itinerary for my car trip for the weekend and book tickets”.
Consider the difference:
Traditional Search (Tool): "Best hiking trails near Boulder, CO.". Returns links to hiking blogs, AllTrails, and Wikipedia. You click, read, compare.
AI Agent Search: "Plan a moderate 3-day hiking trip near Boulder, CO, for a family with two teenagers next August. Include gear recommendations, dog-friendly options, and suitable accommodation within a 30-minute drive that has good reviews and a pool." The AI agent would then potentially:
- Generate a detailed itinerary with specific trail suggestions.
- Provide a curated list of gear with links to purchase from reputable retailers.
- Filter accommodations based on specified criteria, cross-referencing reviews.
- Even potentially book reservations or generate a packing list, all within the search interface
Just imagine, you don’t have to press 10 buttons in your app to make a transaction, save it in your phone. The agent will purchase it and make the payment for you. Take it another step further, your Google Home Mini will now be able to do Macro tasks, and you don’t have to be constantly tied to your phone screen, unless you have the habit of doom scrolling, of course.
I have my own apprehensions about AI and what it would do to people, but I am definitely on board for this one.
And what about you? Do you view AI as friendly like Jarvis, or more like Terminator? Feel free to reach out and share your thoughts on Email, Reddit, or LinkedIn.