Edu

• CS is no longer the top major at most universities (low confidence in this as a prediction, but it’s inspired by the fact that software is increasingly easy to learn independently and increasingly easy to create, students are better off studying something else and learning CS on their own)

Crypto

• Crypto remains niche — increasingly seen as a community that entertains itself — not a bad thing, it’s just a new market entirely rather than a technology that is going to consume the world. Too niche — too much of a “social software”. There are interesting benefits but you need to desire the benefits. No meaningful application is built in 2024 despite positive market movement.

VR/AR

• Apple Vision is too expensive for any meaningful adoption despite being a really cool technology. Media and entertainment remain the primary use cases for AR/VR/MR.

ML

• In the same way that Google pays apple billions of dollars to be the default search engine on iOS devices, some LLM provider (OpenAI, anthropic etc) will pay large sums of money to be the “default LLM” of some hardware device. LLMs are mostly commoditized the same way search engines are mostly commoditized. Sure, there are differences in their performance / quality of results, but for the most part you could use any of the options interchangeably and be mostly satisfied with the results. Competing at the foundation model level seems like a fools errand so long as there are sufficient options on the market.
• Deepfakes attempt to interfere with election but are not quite at the point where it very meaningfully interferes with election. People will be able to identify deepfakes, especially with tools like twitter community notes. If anything those affected will be fringe (older gen etc.) Another possible path is deepfaking “ordinary” people — not faking celebs (which are way too easy to debunk) but faking regular old people with alt-right beliefs to infuriate political sides and incite infighting.
• Apple announces a local first LLM / AI integration (separate from Siri). Consumer electronics that can run GPT-4 tier or higher inference do make it to market, although the form factor is not yet humane or tab. People still like their phones. Photography matters too much. Humane et al do not feel like they can replace phones.
• We do not get AGI, “AI summer” cools down and we realize that state of the art boom in 22/23 is more asymptotic than exponential.

Web Dev

• react is still the main framework and there isn’t any meaningful change. All of the front-end framework competitors on the market are great for folks who like to invent problems for themselves to solve but anyone working on real products is going to use the framework with the most adoption and strongest community (react).
• “browsertech” becomes the next big crazy in web / software. The “figmaization” of software

Society

• Privacy becomes more important / salient in people’s minds. People are more aware of the bias and manipulation present in social media in particular.
• Big tech is increasingly seen as a negative influence, more people leave social media.
• remote work (sadly) dies down, although folks realize that being in “hubs” does not really matter much and is more important for younger folks who want to feel like they are in the center. Physically working together helps, but great work can be done anywhere.

Grab bag

• The vibe is positive