Piecing all the evidence together, allows us to recreate the chain of events that led to the pandemic. It perfectly fits the evidence, and does not require making any unusual assumptions:
Despite rejection of the DEFUSE grant, WIV undertook similar research, relying on their existing collection capabilities and virus samples.
Specifically:
- Identify SARS-like viruses 10–25% different from SARS-1.
- Screen those with a good match to the human ACE2 receptor.
- Insert an FCS at the S1/S2 junction.
- The modified virus was cultured and used to infect laboratory animals, as described in the proposal.
- This was done under insufficient safety conditions — BSL-2
- A lab worker became infected, possibly without realizing it, and transmitted the virus to others, probably in late October 2019.
- The infection spread through Wuhan, unnoticed on the background of peak flu season, and quickly reached the high-traffic HSM
- There, the crowding, poor ventilation, and wet surfaces caused it to rapidly spread among the employees.
- A cluster of cases in a market with wildlife was hard to ignore, given China’s experience with SARS, and the virus was first identified.
1
289
In exact sciences, great weight is given to hypotheses that can predict future observations. When dealing with complex systems, there are no controlled experiments, so accurate predictions are difficult. However, when all the evidence that comes out after the hypotheses were laid out strongly favors one hypothesis, that is a strong indication of its superiority.
So while we were very disappointed that the judges failed to identify the correct hypothesis at the time, the fact that the debate forced each side to lay out their most likely hypothesis provided a way to clearly test which better survives the test of time.
1
309
Interestingly, testimony from Ralph Baric released after the debate confirmed this: one motivation behind the DEFUSE proposal was studying Antibody-Dependent Enhancement (ADE) in feline coronaviruses, some of which also carry the PRRAR sequence.
Baric specifically explained that cleavage efficiency is nonlinear: excessive cleavage can reduce infectivity, which makes the “suboptimal” PRRAR configuration biologically reasonable.
In 2022 a FOIA request revealed early versions of the 2018 DEFUSE grant proposal written in collaboration with WIV. This proposal discussed engineering of SARS-like viruses, and while it wasn’t approved, it provides a glimpse into the kind of work that coronavirologists were interested in pursuing at the time, meaning WIV could very likely execute portions of it independently.
Before this FOIA, zoonotic supporters claimed the work in DEFUSE is too different from SARS-CoV-2: (i) It discussed cleavage sites that are not necessarily an FCS, (ii) It proposed inserting them in the S2 region, rather than the S1/S2 junction, and (iii) it seemed to focus on viruses similar to SARS-1, while SARS-CoV-2 is 24% different.
The full drafts exposed in the FOIA resolved all of these discrepancies: They specifically mentioned inserting an FCS at the S1/S2 junciton, and proposed doing so with viruses that were 10–25% different from SARS-1, because those are more likely to evade immunity for SARS-1.
Source: teapawt.net/ydeigin/status/1965041…
Replying to @MonaRahalkar
Baric gave WIV the roadmap what new SARS-like viruses to look for ("highly variable", i.e. spike 10-25% different from SARS1), and what cool new stuff to try (add an FCS, play with glycans). After that, WIV didn't need him anymore. It was Baric who needed WIV for novel viruses, and he knew they weren't giving him the most interesting ones:
1
345
This alone is enough to undermine the main argument for zoonosis, but we actually have concrete evidence that came out after the debate and directly contradicts the market spillover.
Scientists classified early genomes into two groups: Lineage A, and Lineage B, which descends from A. All cases in the market were of lineage B, which is weird for the initial spillover location - we’d expect it to have the earlier strand - lineage A. Zoonosis supporters relied on a study pointing to two separate spillovers from animals to humans at the market, one for each lineage. Besides the implausibility of two spillovers happening in one location and nowhere else in the world, this study was later found to have multiple errors.
However, the final blow to this hypothesis was the finding of intermediate human variants between A and B, showing they both descended from the same ancestral virus already spreading among humans. This sits well with the lab leak hypothesis: There is one spillover from the lab of lineage A (or its predecessor), which mutated to B and only then reached the market and rapidly spread there.
Source: teapawt.net/ydeigin/status/1782497…
Replying to @ydeigin
79/
That alone severely undermined the double spillover hypothesis. However, what provided dispositive evidence for rejecting it, was confirming the existence of true intermediate early genomes between lineages A and B (called B0 in the quote and figure below):
academic.oup.com/ve/article/…
3
1
367
The last example is the US intelligence assessment. In the section arguing for zoonosis, the first cluster at HSM isn’t even mentioned as evidence, and they specifically cast doubt on its role.
dni.gov/files/ODNI/documents…
1
424
A classic example is a forensic DNA match. A prosecutor might claim there’s a “1 in a billion” chance that an innocent person’s DNA would match the crime scene, implying near certainty of guilt. But that assumes a coincidence DNA similarity is the only possible explanation. If we truly assume innocence, other explanations (e.g. contamination, lab error, transfer via objects, framing) are far more plausible. Clearly, it isn’t one in a billion.
The same reasoning error occurs with the “1 in 10,000” claim about the market cluster. That number assumes the cluster could only appear there by pure chance, ignoring the many other ways such clustering can occur.
1
545
The key mistake was treating the market cluster as a “1 in 10,000 coincidence.” That number came from the share of Wuhan residents employed at Huanan.
The argument seems compelling: “Of the thousands of possible places, what are the chances that the first cluster would happen to be exactly where wildlife animals were traded?”, but that kind of thinking is actually a common mistake in probabilistic inference.
1
646
As we showed on the Rootclaim blog, the only evidence supporting zoonosis was that the first cluster of cases was identified in the Huanan Seafood Market (HSM), which also had a few wildlife traders.
Scott Alexander’s calculations @slatestarcodex showed that when this evidence is removed, support for a natural origin collapses, and even Scott’s own conclusion flips from 94% zoonosis to 94% lab-leak.
blog.rootclaim.com/covid-ori…
1
977
Interesting updates since the Rootclaim debate on covid origins - lab leak or zoonosis?
@Rootclaim founder @saarwilf, interviewed with Destiny @TheOmniLiberal and discussed how probabilistic inference can help resolve such controversies, and how despite zoonosis winning in the debate, all evidence that came out since then completely contradicts zoonosis.
🧵Thread of the interview’s highlights:
Rootclaim is thrilled to announce the world's first $1 million debate on a pivotal question:
Did the COVID mRNA vaccines save more lives than they killed?
Our opponent, Mr. Steven Kirsch—a prominent vaccine critic—asserts that these vaccines saved few, if any, lives and caused millions of deaths globally. While we disagree with his conclusions, we commend Mr. Kirsch for taking a personal risk and attempting to back his public statements on this important issue.
Debate Details:
• After two years of negotiations, the agreement was signed, including the identities of the two judges.
• One million dollars have been deposited in escrow, and there is no going back.
• The debate will be in writing and video, starting before year's end.
• Texts will be shared publicly to allow anyone to weigh in, ensuring both sides present their best case.
• Videos will be released after the winner is declared to preserve the judges' anonymity.
We invite experts with relevant knowledge to contribute. Please comment or contact us if you'd like to assist, or tag others who might be interested.






