HNHacker News
TopNewBestAskShowJobs

jphenningsen

48 karma · joined May 28, 2019

submissionscomments
jphenningsen··on Our brain is a prediction machine that is always active
Yes, the theory isn't new, but getting experimental evidence is difficult in neuroscience. Here they used GPT-2 for quantitative predictions, which obviously was not available a decade ago. This way they could extend experiments from highly artificial signals to natural language.
jphenningsen··on Bosch develops rapid test for Covid-19
This test is made for use at the point of care, apparently all the operater has to do is insert a swab into a cartridge and put it into the device. [0] Comparing the throughput to fully automated systems used in testing labs [1] (look at this beast!) or qPCR machines in research labs [2] is missing the point, those require high investment or skilled staff. A test that can be run quickly by any doctor would be very useful to rapidly detect new outbreaks in the time following the current lockdowns.

However I suspect that the downside of this device will be the high cost of consumables, as is common for microarray biochip based devices.

[0] https://www.youtube.com/watch?v=jgW2pYmzcco [1] https://diagnostics.roche.com/us/en/products/systems/cobas-8... [2] https://www.bio-rad.com/en-us/product/cfx96-touch-real-time-...

jphenningsen··on Why So Many Londoners Live in 'Two-Up, Two-Down' Housing
The next article in this series gives a great example of desireable flats: How Berlin’s Once-Reviled ‘Rental Barracks’ Became Coveted Urban Housing https://www.citylab.com/design/2020/01/berlin-architecture-h...

The key point seems to be that although these flats were originally undesirable social housing, their spacious, unspecific layouts + somewhat reduced density from WW2 bombing allowed the transformation to the opposite: Expensive gentrified housing in very high demand.

jphenningsen··on Why natural cycles only play small role in rate of global warming
Just from a first look, the caption of the figure references Figure 5 in Haustein et al 2019 [1]. You can see that this figure is way more detailed and includes a confidence interval. This article is clearly not a scientific publication but aimed at a broader audience. As for your other questions about the details of the methods used to gather and analyse this dataset, I am sure you can find plenty of information in the actual paper and references therein.

Generally speaking however let me comment that you will not be able to formulate a meaningful critique of the nitty gritty details you inquire about without the equivalent experience of at least a PhD in the specific subfield.

[1] https://sci-hub.tw/https://doi.org/10.1175/JCLI-D-18-0555.1