Viewing profile — emcq

If we are going to consider using prior runs of the program having the file loaded in RAM by the kernel fair, why stop there? Let's say I create a "cache" where I store the min/mea…

The 1.4s is _after_ having the file loaded into RAM by the kernel. Because this is mostly I/O bound, it's not a fair comparison to skip the read time. If you were running on a M3 m…

This is even slightly more direct: access to WSJ data requires paying LDC for the download, and the pricing varies depending on what institution / license you're from. The cost may…

Be wary of using this model - the licensing of this model seems sketchy. Several of the datasets used for training like WSJ and TED-LIUM have clear non-commercial clauses. I'm not …

Shockingly the human genome itself has not been fully sequenced, despite the human genome project completing years ago [0]. There are difficult to map regions of the genome, some o…

I used to work for a drone company. I made sure to go watch the mechanical testing for prop safety. One of the test objects was a chicken leg, and the prop cut through more than 2m…

Check out sse2neon as a way to substitute your x86 SIMD instrinsics for ARM neon. Perf was good enough to ship in one of the past projects I worked with.

This article is quoting employees conversations from 2016. It wasn't until late 2016 that Google published research on wide and deep recommender systems. I'd be interested to see i…

You can get all this functionality and more working beautifully with Android with Garmin watches. Mine lasts for many days on a charge (perhaps a week without activity tracking), h…

Don't forget Ron Graham who cowrote magical mathematics!

This is easy to resolve by staggering doses which I know at least other hospitals are doing.

Elon's was quoted saying "you have a forest of redwoods and the little trees cant grow." While pointing his finger at California, I can't help but think this really reflects back o…

It feels like we are seeing the result of diminishing returns with technology advancement. To make a product that Apple believes is significantly better than the competition they h…

Google colab is a much better tool than a new computer for getting into ML. It's free, requires no dependency management, easy to share, and has tons of example notebooks you can r…

You can make a model of any size with deep learning. If your concerns are about over fitting there are lots of regularization techniques used in practice like dropout, weight decay…

A better rule of thumb is miles because hours don't reflect the amount of energy put into your drive train. Typical recommendations are every 500-750miles for a mountain bike. For …

You can't really make all things be equal and get increased efficiency and longetivity for most cases. The trade-off are explicitly made to have reduced longevity to have better ef…

Years ago I worked at a small, like 5-7 person startup in soma where we were surprised to find Rupert himself personally come by, later send some chiefs to do due diligence, and cu…

Blaine Washington is right along the Canadian border and next to many islands. The islands are sparsely inhabited and infrequently visited by humans it makes me wonder if maybe the…

A common approach in rendering engines to convert screen space coordinates to objects is to render a second image with light and shadow disabled where the color uniquely maps to an…

The most accurate description is probably the authors built a hearing aid from the 1960s with only $1 today [0]. Its worn around the neck, doesn't have any fitting, doesn't appear …

Both https://audatic.ai/ and https://whisper.ai/ are taking the ML approach for speech enhancement, and the large companies are starting to get on board. It's the future!

Unfortunately live listen is more like a replacement for remote microphones, not the hearing aids themselves. It adds around 70ms latency with a bluetooth headset. This is enough t…