Threads
Every comment on the site, newest first.
It is in there — line 14, the $22 board. The expensive part historically was the Peltier driver and that is what got cheap.
Anyone else notice that the reproducibility argument always ends with "we need better metadata" and never with "we need to fund replication"?
Biosafety take: none of this is a problem at BSL-1 with non-pathogenic chassis. The failure mode in community labs is almost never the organism, it is the chemical waste.
The ramp rate claim is doing a lot of work here — is that block temperature or sample temperature? Those diverge badly above 2 °C/s with thin-wall tubes.
Having done the 50k L version: foam control alone eats the margin they are projecting.
Strongly agree. Our incident log is 90% solvent handling and 10% sharps. Zero biological.
This is the third time this month someone has posted a "cheap X" build where the cost excludes the thing that actually costs money. Where is the temperature controller in that BOM?
Add: get the material transfer agreement signed before shipping, not after. Customs asks for paperwork, not for vibes.
Because metadata is cheap to advocate for and replication is not.
Sample temperature, measured with a thermocouple in a dummy tube of mineral oil. Plot is in the repo under docs/thermal.
Off-target analysis with GUIDE-seq only tells you about the sites you could capture. I would want long-read whole-genome on at least a few animals before "no detectable" goes in a title.
Fair. The preprint does have ONT data in supplementary S4, though the coverage is thin.
Screening baseline is good but voluntary. Voluntary regimes work until one provider decides growth matters more.
We run exactly this curriculum. Two years, zero incidents, several very good student questions we could not answer.
The cost curve here is real but the pilot scale is 500 L. Everything changes at 50,000 L and the paper does not pretend otherwise, to its credit.
It is in there — line 14, the $22 board. The expensive part historically was the Peltier driver and that is what got cheap.
Started my lab entirely on auction equipment. The trick is buying from a closing pharma site rather than a university surplus sale — pharma gear comes with service records.
The review actually proposes three. Section 4. They are weak, but they exist.
Rewriting firmware in Rust is fun but the actual reliability win here was the watchdog, which you could have added in C.
This is the third time this month someone has posted a "cheap X" build where the cost excludes the thing that actually costs money. Where is the temperature controller in that BOM?
This matches my experience. Also: never buy a -80 freezer sight unseen.
Open datasets like this are the single highest-leverage thing a lab can publish. Papers age, data compounds.
I would like to see the word "intelligence" retired from this entire subfield until someone proposes a falsifiable behavioural test.
The teardown is beautiful. Note the flow cell is still proprietary, so "open" has a ceiling here.
Trehalose stabilisation is old but the field-deployment data is what makes this worth reading. 40 °C for six weeks is a real number.
We got a term sheet last year that explicitly cut the CMC budget to make the burn look better. Declined.
Trehalose stabilisation is old but the field-deployment data is what makes this worth reading. 40 °C for six weeks is a real number.
Open datasets like this are the single highest-leverage thing a lab can publish. Papers age, data compounds.
Counterpoint: undocumented data compounds confusion. The README here is genuinely good, which is why it works.
We ran this protocol in our community lab last month. Two of five transformations failed until we dropped the heat shock to 42 s. Worth adding a note.