drzoltar

A better example is aviation, especially commercial. Planes come with Pilot Operating Handbooks (POHs) that contain checklists specific to the aircraft for preflight, inspection and failure scenarios (eg loss of engine power). Manufacturers periodically update these by issuing advisories and addendums to POHs. The FAA also publishes such advisories.

There doesn’t seem to be much evidence that members of congress outperform the market [0], let alone the sector that investments are made in. Aside from a few suspiciously timed trades (eg short sells just before the Pandemic), is there any strong evidence in favor of Congress outperforming the market?

[0]: https://www.nber.org/papers/w26975?utm_campaign=ntwh&utm_med...

Surely there must also be Math Gremlins, who are chipping away at its foundations. CF Gödel and Chaitin’s Incompleteness Theorems.

Didn’t know this happens outdoors. Does the CO2 rise well above the surface though? I’ve heard there are “Dog Caves” [1] around the world where CO2 is low in the cave and will kill short animals but spare adults.

[1] https://www.amusingplanet.com/2018/03/why-dogs-cant-stand-ca...

The other issue with these hashes is non-robustness to adversarial attacks. Simply rotating the image by a few degrees, or slightly translating/shearing it will move the hash well outside the threshold. The only way to combat this would be to use a face bounding box algorithm to somehow manually realign the image.

Dumb question: why can’t crypto currencies and exchanges place the ransom tokens on some kind of blocklist, thereby forever tainting those coins? As I understand, the rise of “privacy wallets” has greatly increased the anonymity of such transactions. But, at the end of the day, don’t we always have a ledger of the coin ids? I’m curious how the coins actually get laundered back into cash.

The time has finally come to invest in RNA malware detection.

In all seriousness though, would it be easy to check these vaccines for authenticity via some simple PCR test?

Which organizations changed their guidelines?

Which other paid news publications do you prefer over the NYT?

In California, the main issue seems to be that there’s no money in these charging networks. Most non car company chargers I’ve seen are laughably expensive (sometimes exceeding the price of gas per mile) and slow. Despite being so expensive, it’s hard for me to believe that they make more than $10/day. It’s a few dollars per hour to charge, and most of the time they’re empty. And that excludes the maintenance cost. So you end up with many broken chargers. I’m guessing there’s some sort of state subsidy that helps establish these chargers in parking lots? It all feels like a money laundering operation.

I’m worried what will happen down the road. Robinhood is likely getting super valuable data on this. Combine that with correlating post activity on Reddit and you have a recipe for obliterating the arbitrage of retail investors. These hedge funds will also buy the data that Robinhood freely sells [1]. Sure, squeezes like this will still happen every so often, but on average the hedge funds will win more in the long run.

So, I think Robinhood needs to get some competitors, who charge $1-2 per month, while enforcing data privacy.

This all feels like a classic Black Swan at work, and I guess the main question is if you believe that the long term expected return justifies such risky short sells.

[1]: https://www.jacobinmag.com/2021/01/trading-app-robinhood-inv...

FWIW, general profanity detection is a highly nontrivial problem. It’s true that such subword profanity filters aren’t that great, but slightly more sophisticated ones (eg whole word matching or n-grams) tend to have relatively good precision. You could train a fancy neural network, but the overall return on precision and recall tends to be not that great (compared to the exponential change in speed and cost). The problem almost always crops up in out-of-distribution sentences (such as “bone” at a paleontology conference).

It seems like the quantity of the dose, location of exposure and duration of exposure can drastically change the outcome [1]. The woman who sprayed Novichok-laced perfume on herself died, but the man got it on his palms and immediately wiped it off, thereby surviving.

[1] https://arstechnica.com/science/2020/09/novichok-how-are-vic...

There’s a lot of false equivalence in this post. The message seems to be “if you want to build amazing new things, don’t A/B test.” This is vacuously true. You certainly won’t discover fusion, or make the next iCar through rote testing.

What’s hidden though is that out of the thousands of companies that will try, only a few will succeed. The rest will burn through millions and have nothing to show for it. A/B tests are done because they work in the short term, providing runway for the long term. It’s true that many companies abuse these tests and heavily lean on incremental improvements. Management can become complacent, spineless and rudderless, thereby refusing to make hard choices that go against the data. But, it keeps volatility down and appeases the overlords holding shares and vested interest in the company.

It does make wonder though how, despite being so secretive, Apple got a lot of things right with the iPhone and iPod. So whatever their process was for internal focus groups is definitely worth exploring. We’ve all heard how Jobs was insanely attentive to details, but I still wonder where the overall cohesion of design came from. No one person could have possibly gotten all of that right.

In my workflow, one reason I prefer VSCode over Sublime is for dealing with large (10mb+) files. Sublime takes upwards of a minute to open such files but VSCode is nearly instant.

For smaller files, Sublime is great.

SNPedia has a paid service that can do it for you:

https://promethease.com/

I personally didn’t trust them with my data so instead you can just search for specific SNPs from your DNA file. For example, the coffee one is this one:

https://www.snpedia.com/index.php/Rs2472297

So you search your dna file for rs2472297 and then check which combination you have, e.g., you see “C T” which is slightly associated with coffee consumption. You can probably write or find a script that does this for you.

I recently downloaded my Ancestry.com DNA data, and started looking at it with the help of SNPedia:

https://www.SNPedia.com

One of the neat things I found is that I have genetic markers for increased coffee consumption:

https://www.snpedia.com/index.php/Coffee

I drink at least a few cups a day and so does my dad, so confirmation bias does the rest.

That makes me wonder if drinking coffee (amoungst other foods) is truly causal in decreasing prostate cancer, as opposed to heavy coffee drinkers having other correlated biomarkers (in particular ethnic correlations) that help dodge it.

Anyways, I’m not trying to disparage this study, I just think it’s neat that we’re somewhat predisposed to like certain foods from birth.

My understanding is that high SWE salaries require the confluence of two factors:

- Competition for talent.

- (Potential) Output value.

I think a very rough rule is that at a FAANG-like company, a really good engineer can be delivering at least $1M in revenue/value, per year.

So the only logical conclusion seems to be that there’s too much supply of qualified engineers. But then that leads me to believe that many big European tech companies are just piling up money in a vault somewhere.

So my question is: who gets that vault money?

I think you’re referring to https://acloudguru.com/

If yes, they’re quite good!

Are there any other shows that are on par with The Wire in terms of complexity, realism and visual story telling? I haven’t seen anything else that even comes close.

For a stark reminder of the inequalities in Covid impact, there’s a nice map here detailing Los Angeles rates by subregion:

https://products.xtown.la/coronavirus

For non-LA natives, the west side (near the coast) is considerably more affluent than the rest, and sees much lower rates. Even places like Palms and Brentwood which are packed with apartments seem to be doing alright relative to the rest.

Try posting it on Crossvalidated: https://stats.stackexchange.com/

Or on Math Stackexchange: https://math.stackexchange.com/

Could it be a misunderstanding of the outcome, or perhaps bias in Facebooks algorithms? The best predictor of your target is: your target. In this case it’s predicting who will buy your product. Using a proxy such as income and education is a good start but surely cannot be the best approach because it artificially limits the search space. The best target is some ML-determined combination of features. My guess is that for niche items such as expensive watches and cars, Facebook’s targeting may be too liberal with lower income earners. But I think this the result of being very good at targeting the general audience for medium-low priced goods and services.

I’m a bit confused why we are assuming the low performance is the fault ML. I would think instead that this low performance is deliberate because it locally optimizes Facebooks KPIs.

My understanding is that Facebook has a lot of data on the elasticity of ad buyers. If Facebook were to have even more precise algorithms, then they would likely charge even more money for impressions and clicks. Presumably buyers may be reticent to pay even higher prices.

Alternatively more precise targeting would highly bias the money people spend toward larger companies that can afford better keywords.

Could you elaborate on the suite of tools you used to do this kind of network traffic analysis?

I have a different theory. I’ve noticed such (and similar) ads appearing a lot on some news sites. I think these sites use sexy ads to drive future traffic.

In other words, these ads lead to higher future retention by subliminally pulling users back for more content. Perhaps they help to amplify and complete a dopamine loop that occurs from reading bleak content such as war and COVID, or low-stimulating content such a recipes, or making us stomach emotionally-charged political arguments. They make us come back for more.

It should be easy to do. You look at your user retention in control vs test groups that don’t show and show these ads, and track retention. At the (small?) price of a worse conversion rate, you retain your users better with such ads. If someone has a nice psychology source that could corroborate this, I would be thrilled.

My impression is that places like DeepMind are more of a social good than a profit seeker. Most universities simply cannot afford such big and expensive AI. So I suspect that DeepMind is Google’s way of expanding the frontier of AI knowledge through charitable donations. It’s ultimately good PR for them and maybe influences businesses to pick Google Cloud Services for brand appeal. Also I’m guessing since it’s research, they get a huge tax write-off.

Perhaps they will deliver once-in-decade unicorn breakthroughs that can ultimately be monetized? Maybe Google’s next phase will be AI generated content. With that they can completely cut out other content creators, and thereby collect affiliate cash through sales. For example, generating those “Top 10 Holiday Toys” guides automatically.

Machine Learning in Clojure reminds me of Yann LeCun’s ML course from 2010, where we used an adorable language called Lush:

http://lush.sourceforge.net/

which I suppose can best be described as Lisp and Python having a baby. It was immense fun to code neural networks from scratch in it. I hope Clojure can find a bigger place in the world of ML.