Author: Zuo Ye
The black box of large models is maddening, while blockchain hopes to enable transparent white boxes in scientific research.
In 1943, the owner of quantum state cats, Schrödinger, gave a difficult lecture in Dublin, discussing the relationship between atoms, life, and cells from the perspective of statistical physics. At that time, across the ocean, the young Watson was only 15 years old, but already a freshman at the University of Chicago.
After reading Schrödinger's lecture, Watson wrote 'What is Life,' determining that genetics would be his lifelong ambition.
Ten years later, when Watson, who had obtained his PhD, proposed the double helix structure of DNA, the 25-year-old had already locked in the Nobel Prize in advance.
Grafting, cloning to gene editing.
There are two trees in front of my house, one is a jujube tree, and the other is also a jujube tree.
Anyone who has been to middle school knows that genes are information fragments of DNA, like the 'function body' in code, which is the most basic functional implementation. DNA is like an instance module, RNA is like routing and communication functions that transmit gene information to specific targets.
Watson discovered the structure of DNA, but humanity did not know how to utilize it. It's like we know about Schrödinger's cat in a quantum state; while the cat is easy to find, quantum communication will take many more years.
At least Watson was luckier than Schrödinger. In the summer of 2012, Emmanuelle Charpentier and Jennifer Doudna discovered that CRISPR sequences and Cas proteins could be combined to artificially cut specific sequences and insert their desired information fragments, ultimately utilizing the body's repair mechanisms to unknowingly complete the process of grafting.
It really resembles horticultural pruning, where visible branches connect with each other. Without understanding the biological mechanisms, we can still know the matching relationships between different plants—just through continuous experimentation.
Experiments can continue, so can cloning. Cell nuclei and cytoplasm can also be separated and 'connected.' Through continuous experimentation, cloning can achieve the marvelous effects of isomerism, just as mentioned in 'What is Life.'
Gene editing is not mysterious; it's a further step from cloning, a further micro-scale from size, just as the atomic perspective of life is merely the irreversible process of thermal motion's ultimate stillness, like time, which may stretch or compress, but is forever irreversible.
Humans can graft fruit trees, humans can clone animals, so can humans edit humans?
In 2018, the crazy scientist He Jiankui became either Eve or that snake, editing the genes of a pair of embryos from parents with HIV. Humanity thus opened Pandora's box. While cloned animals can be humanely destroyed, is a genetically edited human still human?
Image description: CRISPR-Cas9 working principle, image source: @zuoyeweb3
However, the in-depth exploration at the genetic level presents a fatal temptation for certain populations—longevity. Finding the gene fragments that affect lifespan, like a hacker from General Jin, modifying its value from 100 to ♾️, even adding a 0 would be enough.
In 2023, Paradigm co-founder Fred Ehrsam decided to leave the crypto industry and establish a biological research company called Nudge. Coincidentally, Fred is also a co-founder of Coinbase, which transitioned into crypto VC after the company went public in 2017.
Also in 2017, Paul Kohlhaas joined Consensys as the head of BD, but left to start his own venture a year later. Why not do something more interesting with blockchain?
For example, in scientific research, Molecule was established in 2018 as one of the earlier explorations of the intersection of blockchain and research, especially in biological studies. Meanwhile, AlphaGo's parent company DeepMind's life science model AlphaFold was released in 2016, demonstrating its power in protein structure observation.
In 2020, AlphaFold2 successfully solved the protein folding problem, and the 25-year-old Watson booked his Nobel Prize. This time, the 4-year-old AlphaFold2 is also expected to secure 50% of the 2024 Nobel Prize in Chemistry.
Fred's career switch in 2023 is not considered early. As early as 2020, another founder of Coinbase, Armstrong, initiated the establishment of ResearchHub, aiming to deconstruct the institutional research process from universities to papers to funding, introducing incentive mechanisms that allow universities to dominate titles, publishers to profit, and to shift the three mountains of grant applications off scholars' backs.
Especially since the submission fees for scholars to publishers are self-borne, but the reviewers selected by publishers often work without compensation, with only the publishers profiting from the difference.
All elements are coming together; AI, scientific research, and papers are converging towards life sciences. The 21st century is indeed the century of biology.
The crypto world offers immortality, first seeking the elixir of life.
Decentralized scientific research (DeSci) aims at pharmaceutical research in the life sciences.
DeSci is a crypto version of the AI4Sci movement, but highly focused on AI, life sciences, and new drug development. Perhaps it has also taken a 'detour' through meme culture; remember Paul Kohlhaas's Molecule? In 2022, it even received investment from Balaji. No one can resist the temptation of longevity.
Furthermore, in 2022, Paul Kohlhaas established Bio Protocol, starting to develop products that would allow crypto big shots to live longer, with multiple sub-DAOs that address the scientific mysteries of life from male head to toe.
In 2024, the 'reborn' CZ appeared at the Bangkok DeSci Day with Vitalik, where the young V recommended the supplement VD001 from Vita DAO under Bio Protocol to the older CZ.
Then, Bio successfully received investment from CZ's YZi, the token smoothly headed to Binance, and Paul Kohlhaas was quite entrepreneurial, keeping up with the times by imitating PumpFun's Pump Science. Is there a future for combining memes with scientific research?
However, after the surge in Bio, dissatisfaction with delivering results followed. In the traditional research field, developing a new drug often costs over 1 billion USD and takes years or even decades, while the secondary market of Bio cannot wait even five minutes. Taking the money without pulling the market is a sin when it comes to real scientific research.
The story doesn't end here, as the wave of Agents has arrived. AI Agents really have the potential to change the efficiency of scientific research. What's more interesting is that ResearchHub received a $2 million investment in February 2025, and publications by DeSci's Agents are also being reviewed.
In August 2025, Bio Protocol launched the V2 plan, creating a brand new Launchpad, BioXP points program, and BioAgents, still crafted with ElizaOS, once again keeping up with the times.
In just 7 days, over 100 million BIO tokens have been staked; however, on August 7 alone, 80 million tokens poured in, so there are still some issues with data accuracy. Besides that, the economics designed in the V2 plan are more reasonable.
Small market value avoids sell pressure, encourages continuous sponsorship of projects.
Image description: $BIO staking data, image source: @cl2pp
However, the progress represented by Bio Protocol is slower than that of AI4Sci. AlphaFold had already open-sourced its database in 2021 and has so far only published 200 million protein structures, covering basically all known species.
Feeling significantly behind in progress, Bio Protocol has always hoped that the FDA would publish or integrate large pharmaceutical companies' accumulated data to accelerate open scientific research.
In addition, Bio V2 is expected to promote the launch of multiple new drugs in the UAE, greatly shortening the traditional R&D process, while the relaxed human experimentation restrictions in the Middle East will also accelerate life sciences research. It remains to be seen whether it's He Jiankui or Watson.
Conclusion
The performance of GPT-5 is disappointing, but in niche fields like medicine and scientific research, we can still patiently await the blooming of Scaling Law. The data potential in these high-value areas has yet to be thoroughly explored, and any progress could lead to a significant enhancement in human cognition.
In the field of life sciences, Silicon Valley's Colossal ancient species revival plan has been advancing, using CRISPR-Cas9 technology, such as creating a 'Mammoth Mouse' by combining mammoths and mice, or cultivating a pure white giant dire wolf from ancient giant wolves.
Perhaps one day, humanity will evolve; perhaps one day, humanity will perish.
