Sam Altman, too many women is no woman at all, my son; the end is loneliness

The debate on artificial intelligence has long revolved around "how smart" it is. In recent weeks, however, it has evolved into a more critical question: Where is AI settling, at what layer is it operating, and what tasks is it taking over? Recent developments in the agenda are now providing clearer answers to this question.

BROWSER LAYER: THE RIGHT PLACE, THE RIGHT MOVE 

By adding agentic AI capabilities to its Chrome browser, Google is transforming AI from a mere assistant that provides suggestions into an actor that performs tasks directly on behalf of the user. Tasks such as buying tickets, filling out forms, and completing multi-step web operations are now becoming possible at the browser level.

This move is the most accurate AI placement Google has made in recent times. Because the destination for this business was clear from the beginning. AI will not flow within applications, but through browsers. The browser is already the de facto operating system of the internet. Google is finally accepting this and acting accordingly.

At this point, Perplexity needs to be placed in a separate category. It is moving along a quiet, unassuming, but extremely correct path. It treats the browser as a layer that thinks, decides, and establishes context. The real race is here.

On the Apple side, the picture is not very bright. It has fallen incredibly behind in AI. After all this time, it is trying to catch up in every field at once, but it has not yet produced a concrete success in any of them. Even presenting ChatGPT integration by routing it through Siri has been recorded as one of the strangest AI uses in the world. It is not hard to think that Tim Cook has completed his days in this game. If you ask me, this had already happened.

PROACTIVE AI: POWERFUL, FAST, AND DANGEROUS 

Proactive AI systems like Moltbot are abandoning the classic "ask-answer" model and moving to a structure that takes initiative. These systems do not wait for commands from the user; they send emails, make reservations, and organize tasks on their own.

The growth rate of such structures is remarkable. The fact that a startup we heard about on Thursday creates its own social media presence by Monday shows how aggressively the field is progressing.

However, there are two serious security problems here. The first is very clear: You are entrusting a computer to AI. Inside that computer are all your information, your assets, and your digital life. The second problem is even more critical. The possibility of someone accessing this system over the internet and, worse, the AI controlling the computer being hijacked. Until these risks are resolved, the widespread adoption of such systems could lead to serious consequences.

OPENAI, PRISM, AND THE ''BEING EVERYWHERE'' DISEASE

One of OpenAI's latest moves is Prism, a special workspace for scientists. The goal is clear: To offer a vertical AI product for academic paper writing, collaboration, and scientific production.

Prism is important for OpenAI because it shows for the first time that they are moving away from the "everything for everyone" approach. At least on paper. However, it also reveals the following: OpenAI is trying to be a leader in every field at the same time.

Almost every week, a new product, a new model, or a new breakthrough is announced. Academia, software, video, audio, agents, hardware... The list goes on. Here, one inevitably recalls that line from Teoman's song: "Too many women is no woman at all, my son; the end is loneliness." Sam Altman may not have heard this song.

The problem with Prism specifically is this: Academia requires depth, not speed. It is expected that AI supports the thought without distorting it, rather than writing the text. OpenAI's effort to catch up everywhere at once increases the risk of products like Prism remaining unfocused. A lot is being done, but it is not yet clear which one is truly at the center.

MICROSOFT: BIG GOALS, LIMITED RESULTS

Microsoft's AI hardware and infrastructure investments, especially the moves it has made on the inference side, look meaningful on paper. However, when Microsoft is involved, a natural caution arises.

The issue is not bad intentions. The real problem is that Microsoft often fails to achieve the successes it targets. The strategy may be correct, but the execution and impact side often fall short of expectations. For this reason, it is understandable to approach every new Microsoft move with some anxiety.

SCIENCE IS ACCELERATING, BUT PLEASE, NOT ASTROLOGY

Last week, there were major AI developments, especially in the field of health. The same momentum is clearly seen in astronomy. Archive data from the Hubble Space Telescope spanning decades was re-scanned with AI, and thousands of celestial bodies that had not been noticed before were uncovered. This discovery, made by re-reading existing data without building a new telescope, shows the true role of AI in science.

Here, AI does not theorize, it does not prophesy. It only makes sense of data piles of a magnitude that humans cannot keep up with. For this reason, astronomy is one of the fields that benefits most from AI. Astrology, on the other hand, does not correspond to any science. The fact that we still have to make this distinction is thought-provoking in itself.

DIGITAL TWINS: THE NEXT THRESHOLD 

The reason NVIDIA is developing the Earth-2 platform is much more than just showing the weather. Earth-2 aims to create a digital copy of the planet, that is, to simulate the climate, atmosphere, and physical processes with AI. The goal is to produce faster, cheaper, and higher-resolution forecasts.

This approach brings the concept of digital twins back to the agenda. While we talk about digital copies of cities, infrastructure, and climate systems, we will very soon start talking about our individual digital twins. This is not just a technical development; it is a threshold that will challenge the concepts of identity, responsibility, and representation.

YOUTUBE AND THE DIGITAL TRANSFORMATION CONTRADICTION 

YouTube has launched a harsh cleanup against large-scale AI content farms. Channels that have reached billions of views and have millions of subscribers have been shut down. The reason is clear: Automatic and spam content that does not produce value.

However, there is a serious contradiction here. Google, the owner of YouTube, is a company that surpassed Yahoo in search engines years ago thanks to automation and digital transformation. Today, on the YouTube side, there is fierce resistance against a form of production that has emerged through digital transformation.

For years, YouTube was a platform that spurred content creators to produce more and faster. Now, it is cutting off one of the areas that could bring it the most content. If content is being watched, the platform should not object to that content.

The real problem is the uncontrolled swelling of the long tail. If I were the CEO of YouTube, I wouldn't touch the blockbuster content. I would clean up the long-tail content that gets very few views. I would shorten the tail, not the body.