higginsc

same! I was excited to see something so esoteric from my grad school days come up

Check out SSCD from Meta https://github.com/facebookresearch/sscd-copy-detection

(full disclosure: I am an employee who has worked with this team before internally, but I was not associated with this paper)

Number of times I've had a true positive alert from Apple's new device tracking notifications: 0 Number of times I've had false positives due to intermittent connectivity issues with my airpods or from my partner using mine: over a dozen. I'm this close to disabling all of these features

Ha! A friend sent me this comment when he recognized this project. Unless there happens to be another firm who did the exact same thing we did, I was a part of this project (see this blog post https://www.svds.com/introduction-to-trainspotting/).

You misunderstood the point of the presentation. The company was a consulting firm that specialized in data science and engineering. Our clients wanted to kick the tires and see what our technical chops were before hiring us but they didn't want to let us use their proprietary and confidential data for our own tech demos.

We didn't want to just use the same open source datasets everyone else did, so we got to thinking about novel datasets we could create that might have applications for industries we sold our services to. From this, the Trainspotting project was born.

Many of us commuted via the Caltrain, which was right next to our office, and we were frequently frustrated with the unreliability (this was in ~2016 or so when car and pedestrian strikes were happening seemingly every week), so we made an app that tried to provide more accurate scheduling.

We used the official API for station:train arrival times, but we found that it was unreliable, so we wanted some ground truth data on whether a train was passing. Since our office was right next to the Castro MTV station, I had the idea to use a microphone (attached to a raspberry pi) to just listen for when the train went by. In addition to ground-truth data for validating arrival times, this gave us a chance to show off some IoT applications. It actually worked pretty well, but it had false positives (e.g. the garbage truck would set it off). So we added a camera.

We pointed it at the tracks and started streaming data off of it. At first we used very simple techniques, processing the raw stream on-device with classic computer vision algos (e.g. Haar cascades) in openCV. We discovered that the VTA, which had a track parallel to the Caltrain and was "behind" the Caltrain in our camera's shot, could cause false positives. Gradually we used more and more complex techniques like deep learning, but the raspberry pi couldn't handle it (IIRC it could only process a single frame in like 6 seconds). So we used a two-stage validation whereby the simpler, faster detectors that could run on the raw stream in real time detected a positive and then we'd send a single frame to run deep learning.

TL,DR: The whole point was to be a tech demo, not to gauge the speed. The trains were either stopping or pulling out of the station, so speed would have been useless.

Instagram doesn't do facial recognition though, so this seems pretty baseless.

I'm surprised and skeptical about some of this. In particular:

  We also measured networks across more than 90,000 Microsoft employees in the United States. Frankly, we expected to see them shrink significantly, given the rapid shifts in environment, daytime rhythms, and personal responsibilities. Instead, we discovered that most employees maintained their existing connections. Even more encouraging, most people’s network size increased. We had assumed that in a time of crisis, employees might strengthen networks within their own work groups in an insular way. In fact, we saw network growth not only within existing work groups but also across different groups, indicating that to adapt and thrive teams sought to build bridges.

I wonder what evidence they actually used to back up this statement. I suspect the effect is actually in the opposite (and more intuitive) direction, and that their finding is actually a result of the McNamara fallacy.

[source: PhD in drug design specializing in G-protein coupled receptor pharmacology]

Drugs like LSD form non-covalent bonds with receptors. In a macro-world analogy, think of your hand sticking to a syrup-covered fork vs. a covalent bond being your hand stuck to a super-glue covered fork. Different drugs have different levels of "stickiness" (called affinity) for a particular receptor, and LSD has pretty high affinity for its target receptor, 5-HT2a, but it isn't permanently attached. In fact, affinity is defined by relative association vs. dissociation rates of drug-receptor complex.

On a microscopic level, each molecule of LSD is falling in and out of the receptor binding site stochastically. This leaves open the possibility of another drug binding to the same site (called competitive inhibition) when LSD isn't occupying the site. If another drug has a higher affinity, it will occupy the receptor more of the time. When you zoom out and consider the entire set of receptors and drugs, the sum of these individual stochastic events these effects follow characteristic patterns described by the law of mass action.

https://derangedphysiology.com/main/cicm-primary-exam/requir... is a pretty good description

*I'm simplifying parts of this somewhat.

Personally, I think that the effective ban on psychedelic research is a crime against science and mental healthcare - given its immense potential...I am happy to we are slowly starting research into that.

Have you thought about supporting the people who have been laying the groundwork for the current wave of research (e.g. https://www.heffter.org/)? This new crop of research owes a lot to this group's decades long mission to bring psychedelic research back to acceptability.

[no affiliation other than knowing some of the founding board members]

BJ's models are breathtakingly simple and universally applicable. Highly recommended reading

This whole thing seems overblown. The conditions required to induce NDMA accumulation are so far outside the normal bounds as to be farcical. Every drug I’ve ever seen advises you to keep it in a controlled temperature environment.

Incredible. I wish people would take themselves less seriously like this. I would never have read that paper and learned a bit about that space if it hadn't been such an engaging read.

They top out at 15 mph in SF. I wish they went faster. I am a cyclist and usually maintain close to 20 mph (not counting lights). It's kind of maddening not to be able to go 20 or 25 on them.

At a sufficiently large scale, even rare events happen constantly.

Can confirm

No. Did you read the OP's article? I'll excerpt a relevant section for you:

Although the current focus on isolation is often described in the media as “the loneliness epidemic,” Robin Hewings, Director of Campaigns, Policy and Research for the U.K.’s Campaign to End Loneliness, warns that for a subjective, self-reported experience like loneliness, “it’s not very easy to make comparisons across time, and it’s not obvious that it’s getting worse.”

While he acknowledges that aging populations mean that there are likely to be a greater number of elderly people around who are suffering from isolation than in previous decades, he also points out that when you look at the percentages of those affected, the trends are harder to discern. “There was some work done in the late 1940s, which would seem to suggest a not dissimilar level to today. This is right at the speculative end,”

5:33

no

“Because Facebook is already in the hands of over a quarter of the world’s population, it is imperative that Facebook and its partners immediately cease implementation plans until regulators and Congress have an opportunity to examine these risks and take action,” the letter says.

We're scared that you have the ability to disintermediate our control of the financial system just like you did to publishers.

Yep. I recently spent a day driving an S5 and S6, and the non-touchscreen carplay was terrible. I have a touchscreen version in my 2017 CR-V, and it is a joy to use. I read the comments in this thread, and it makes me laugh. So much "TOUCHSCREENS ARE THE DEVIL, NOBS AND BUTTONS ONLY." It just doesn't jive with my experience at all. Maybe it's a generation gap thing? I'm 30 FWIW.

haters gonna hate

Hoping for you too. It's in my mom's side of my family. I'm fortunate that my mom didn't inherit, but it's a large family so many loved ones weren't so lucky. That's why I studied biochemistry and neuroscience and got a PhD in drug discovery. Unfortunately my adolescent passion met the realities of real world of biomedical research, and now I work at FAANG. I still hold out hope that others will find a treatment someday for people like you and my family.

It's a fucking desert.

That is pretty funny. The more you know. I was just a grad student and met Shoichet at conferences.

I have been out of this space for a few years (transitioned to data science from drug discovery), but from my time doing in silico and in vitro work, a major issue with docking was rank ordering. His comments are right on the mark IMO. Especially this paragraph:

Another point is that high-middle-low effort on the D4 case. The binding assay results compared to the docking scores are shown at right. You can see that the number of potent compounds (better than 50% displacement, below that dashed line) decreases as the scores get worse; the lowest bin doesn’t have any at all. But at the same time, there are a few false-negative outliers with binding activity at pretty low scores, and at the other end of the scale, the top three bins look basically undistinguishable. So the broad strokes are there, but the details are of course smeared out a bit.

These methods can filter millions of compounds down to hundreds, but as an academic lab, it's still a herculean effort to synthesize hundreds of compounds. And out of those hundreds, you might get a couple that are active. This study is a combination hard work, yes, but also a lot of money and luck.

That being said, good for the team, and good for science. I have nothing but respect for Shoichet and Roth. Didn't ever cross paths with Irwin.

Yeah, that's a pretty strong signal that your account got used for some nefarious shit.

I've heard this about Instagram, too, which makes sense I guess since they have the same ad infrastructure. One woman I was talking to the other week said her favorite thing about Instagram was the ads. I was like.. "huh."

But the point of this was market research. I highly doubt anybody gives a shit, even the uptight brigade here on hackernews, about whether company X knows I'm using company Y's app. It's not like FB wants to read your private texts. They want to know what apps you're using and how you spend your time so they can make their own products better.

Yep, they do. I know several and have been through part of the interview process (I withdrew because Google takes too long, and I went to FB instead of dealing with their bullshit).

I've been an iPhone user for a decade and am a huge Apple fanboy, but even with that bias, I have to say this is a load of crap. Apple has been adding loads of undeletable bloat since forever. they only really let you start deleting the Apple pre-installed apps very recently. I still have a folder on my devices called "iCrap."

"Yeah, but what really gets under my skin is that every other company out there is okay with producing garbage. Meanwhile, Apple produces things that work comparatively well, doing what you expect and then, claims that reasonable functionality is premium."

If the standard is garbage, then functionality is premium.

I've been waiting for a graph view for Jupyter since it was called IPython. This is great!