@martinkoopman@patrickc The rule you describe doesnโt enable freedom. It only takes away the freedom to take away someone elseโs freedom. That is a generally desirable thing, but it is still the opposite of a freedom.
@patrickc Plausibly, nearly free, nearly unlimited intelligence will equalize people with each other to a greater degree than the internet did.
If learning gives people pleasure, perhaps AI will yield a hedonistic paradise.
@ginas_HR@CNGI_news Itโs an endowment. That means the profits from those expensive hotel stays will help make college a bit more affordable for many students in the future.
Deep in the recesses of @ucsf, for five years and $67 million, @GSK funded some of the hardest CRISPR experiments ever run.
In 2025, Alexander Brown took the helm of the operation and this week, he opened the doors to everyone else as their work with GSK winds down.
The Laboratory for Genomics Research was founded in 2019 by Jennifer Doudna with Jonathan Weissman and Hal Barron, a hybrid lab across @UCBerkeley, @UCSF and GSK built for functional genomics at a scale no lab could reach.
Alex took over in August 2025. He came from industry, doing some of the hardest to engineer CRISPR programs money could pay for: four and a half years at @Synthego building and maintaining its Cell Engineering Services, then leading development at EditCo Bio when that function spun out. At EditCo he brought the first genome-edited human T cells to market.
This week he and his team, Gil Sharon (also ex-Synthego) and Sailaja Peddada, laid out what the LGR will now run for outside groups.
SCREENS
Genome-scale and targeted. Pooled and arrayed. Knockout, interference and activation.
LIBRARIES
Custom pooled and arrayed design and construction. Single and dual guide. Plasmid, glycerol or lentiviral. Built for high uniformity and low skew.
AUTOMATION
BSL-2 automated workcells. High-throughput liquid handling. Automated colony picking. Parallel processing built for scale.
READOUTS
Single-cell genomics and Perturb-seq. Arrayed high-content imaging. Flow and FACS. Multiparametric phenotyping.
DATA SCIENCE
Screen analysis and hit prioritization. Bulk and single-cell analysis. Image analysis. Biological interpretation.
MODELS
Immortalized lines, primary cells, iPSC-derived models and 3D organoids.
What that infrastructure produces: Viviana Salinas-Rios is building over a thousand bespoke editors at the Innovative Genomics Institute to find the most precise one for individual patient variants from the clinic.
Daniel Fletcher ran a genome-wide screen on cell surface crowding, where crowding costs an antibody 10 to 100 fold in binding affinity, and found no master regulator, just dozens of genes nudging it.
Also presenting: James Nuรฑez, Michael McManus, Roberto Zoncu, Han Yang, Anna Molofsky, Martin Kampmann, Zev Gartner, Alanna Schepartz.
@DoudnaJennifer closed the day with @genophoria of @arcinstitute and David Raleigh,
with a quote:
"(The) kind of data you're all generating is incredibly important. It's high quality data, which we all know doesn't always exist."
If you have a screen you cannot run yourself, whether you are in academia, a startup, or a Fortune 500 pharma, they are reachable at:
[email protected]
Thanks to @illumina , @thermofisher and @wilsonsonsini for sponsoring such a wonderful event.
@Lavender1985520 There are at least four types of learners:
1. Doesn't learn
2. Learns from experience
3. Learns from observation
4. Learns by supposition and reasoning
Do you know of any others?