Month in mind: 2025 07

A maniacal sense of urgency

In his book Elon Musk, Walter Isaacson highlights Musk’s relentless emphasis on a “maniacal sense of urgency” as an operating principle. This is my experience in this month. If you want to do something, you have to do it immediately. If you don’t do it immediately, you will never have the chance to do it. This motto was always true, but lately has been more true than before.

Reinventing surfaces

It has been widely argued that modern AI has a capability surplus but an integration deficit. But where can we find integration? Perplexity and OpenAI take the lead to renew the browser war after nearly 8 years. Large models such as Gemini release CLI clients, offering a novel interface than web apps, capturing a different user base.

It seems with a saturation in power, distribution matters more. The surface layer of our apps in 2025 is clearly in flux. As products begin to converge on utility, they can only compete for a more upstream position in the information distribution pipeline, which means rethinking their surface forms just in order to get closer to the users. This is what we observe today playing out across general-purpose utilities such as browsers and chatbot assistants, but also in more fine-grained verticals like therapist AI products.

There can never be a good time

The recent Windsurf drama reminds me of another company named after the chicken and egg paradox. Chegg first went public in Nov 2013 at a price of $12.50. With the pandemic and digital transformation boom, Chegg surged and hit an all-time closing high of >$110 in Feb 2021, and then quickly plunged to lose more than 80% of its value within a year. In the shadow of generative AI, as of 2025 its stock is priced at ~$1, less than 1% of its historical peak.

It seems key developments that dominated Chegg’s rise and fall were not always predictable. In the future, we could only expect increasingly so. How could one have done better? When one acts reactively, less is in one’s control. When one has less control, one can only compensate by being unpredictable. To me, this tension might be the driving force behind the increasing volatility in how AI practitioners make decisions today - and feeding a broader culture of opportunism in the industry in general.

Have a great month!