hang xu

First of all, thank you so much for all the support and feedback on launch day, I really appreciate it.

I just shipped a new update to stillmail.

You can now send a letter to a friend and they can read it instantly through a secure magic link, no sign-up required. The link only opens that specific letter.

I spend most of my time on these two things:
1) observing products on the main page and upvoting/commenting on them

2) talking with people within comments under my forums

The second one is the one where I notice some things that I lack and see possible improvements, namely:

Despite the hype, the frontier of AI might not be where most people are looking.

While Big Tech continues to throw billions at scaling transformer-based LLMs, a recent Delphi study shows that 76% of leading AGI researchers no longer believe models like ChatGPT will meaningfully contribute to superintelligence. The marginal gains are becoming too expensive

GPT-5 is estimated to cost up to $1B to train, yet the improvement is mostly superficial. What we re getting is better packaging, not better reasoning.

Meanwhile, a quiet shift is happening in the background. Research in alternative architectures is starting to show surprising results. Liquid Neural Networks, for example, are achieving task-specific performance far beyond what transformers deliver using a fraction of the resources. Neuromorphic chips are pushing 10x energy efficiency. And miniaturized, domain-specific models are proving far more practical at the edge than scaled-up giants.