You want robots in labs because they will improve scientific productivity. I want robots in labs because they will swamp the supply of cheap labour, uproot the abusive academic hierarchy, and eliminate the overproduction of PhDs. We are not the same. adam.rochussen.xyz/p/robots-…

Alright. Andersen just finished his talk. It was slick and persuasive. If I hadn't asked my question, I'm sure most audience members probably would've come away thinking zoonotic spillover was definitely how COVID started. Brief summary of his case and why I didn't find it convincing: Andersen began his talk by setting out that he wanted to ask how COVID began, but then the entire talk was focused on a different question: where COVID began. This is a rhetorical sleight of hand that I was happy to point out to the audience afterwards. Andersen convincingly showed that early cases of COVID were clustered around the Huanan Wet Market in Wuhan. Interestingly, all of the maps on his slides (taken from a WHO report) had an inset that blocked where the Wuhan Institute of Virology is on the map. So audience members would have had no idea that it is a half-hour drive away. Andersen then showed images that his colleague had taken in 2014 of a racoon dog in a cage. WHO had originally stated that there was no evidence of illegal animal trading at the market. Andersen said that this is a lie and that they covered it up. Interesting that this is an example of a conspiracy theory that he believes in (I think he's right!) He then spent most of the rest of the talk showing spatial maps of the market and showing where SARS-CoV-2 sequences were found—pinning it down to a single cage where a racoon dog was thought to be held based on his friend's photo from 2014 and on mtDNA sequences detected in the same places as the SARS-CoV-2 sequences. His other point was to suggest that there were two zoonotic spillovers because there were two independent COVID lineages sequenced from the market. What Andersen showed convincingly was that the epicentre of the pandemic was probably this wet market. I don't think anyone disputes this fact. Thousands of people pass through that market every day. What Andersen didn't show, but implied and hinted at through innuendo, was that infected animals were the source of the disease. I got to the mic first after the talk and pointed this fact out. He was receptive to it. I mentioned to the audience that we are talking about environmental swabs that were sampled by the Chinese CDC based on where they thought the human cases originated—sampling bias is obviously a potential issue here. Many of the samples were also taken from mobile objects (carts, tools, etc). A key point that refutes the zoonotic spillover case is that nobody has ever found an infected animal. I think audience members were not aware of this fact when I shared it. I also pointed out to the audience that these samples were not simply a binary case of [SARS-CoV-2 + racoon dog DNA found together]. The raw data (which Anderson didn't show) show that there was a mishmash of a bunch of species' mtDNA and the SARS-CoV-2 sequences were *incredibly* sparse. One of the areas was labelled as SARS-CoV-2-positive due to a single molecule of viral RNA among over 200 million reads. And qPCR of the same sample came out negative. This is in the realm of technical false positive for the RNA-seq! There were other things I wanted to say, but others were queued up behind me at the mic, so I had to cede. I finished by saying that none of what Andersen presented is inconsistent with a lab leak: a lab worker could easily have gone to the market after work for dinner, infected some other people there, and then it spread rapidly thereafter. This is what I think really happened. Andersen's view is that the double lineage and the distance between WIV and the market refute this. I disagree with the latter (see map), and the former has been shown to be a statistical artefact (michaelweissman.substack.com…). Sadly I didn't get to tell the audience that the two lineages differ by a couple point mutations and that the original Bayesian phylogenetic analysis was fudged—reanalysis showed that a single original lineage is almost as likely as two distinct lineages. I also never got to mention Jesse Bloom's reanalysis of the same data that Andersen uses (pmc.ncbi.nlm.nih.gov/article…) which didn't bias the mtDNA search to just mammals. Bloom found that the highest correlation of SARS-CoV-2 genomes was with catfish mtDNA. Catfish cannot be infected with SARS-CoV-2. This fact shows clearly that spatial coincidence of these samples does not mean any animal was ever infected. I enjoyed Andersen's talk and I think I offered pretty normal, respectful scientific pushback in the Q+A. Hopefully audience members reconsidered the strength of his claims in light of the points I got to make. It was nice to have a forum to discuss an overly politicised issue in a very scientific, open manner.