georgeek

Same here, from both interns we've had who took a class of hers..

It's eye-opening that the salary of an adjunct faculty member at Cal with 100% appointment with classes taught to 1.5K+ students (CS61A) is less than the tuition fees of 8 students.

When your family and kids appear, work stops being a 0-sum game. Instead, you're constantly arbitraging between it and the rest of your life. The good news is that you can recognise and structure it as a positive or a negative sum game.

Hasselblad has its roots in military optics and space-constrained imaging solutions. They have made many professional surveillance systems over the years with smaller sensors. Hopefully this is not just a slapped logo.

Burgers and sausages are some of the most detached ways to eat meat. While a steak has bones, a burger is much more abstract. There is probably a good anthropological theory behind this.

Which is why you should always separate a transaction into smaller chunks - you could liquidate 20% of a position once per week.

Banks do compete but it's not really an iterated game - a company only goes public once.

Almost everything that one would need for daily subsistence is tied to benchmark future contracts: cattle, soy beans, rice, wheat, sugar, coffee are all traded as futures on exchanges. Most coffee in the world is traded against the Coffee C future benchmark for instance: https://www.theice.com/products/15/Coffee-C-Futures

Giving producers and consumers a way to get a standardized contract at a given point ahead in time usually makes the price of the underlying commodity much more stable.

It has been argued that the way our society operates is based on the concept of futures contracts. It all began with the Dutch tulip futures in the 17th century that stabilized flower prices and made them a viable business: https://en.wikipedia.org/wiki/Tulip_mania

Chronicle Software's OpenHFT stack has a few of their libraries: https://github.com/OpenHFT

Such competitions might have two goals in mind: recruiting and signal diversification. The recruiting angle is obvious.

Any alpha that is not fully correlated to existing alpha is worth its weight in gold for an organization with the size, sophistication and complexity of JS. That's part of the reason why efforts such as 2Sigma's Alpha Studio exist: https://alphastudio.com/

The effect will only be seen in the future, smoothed over a few years.

C&JC figured a way to finance their projects through a few big landmark donations along with a proto Kickstarter campaign. They would sell numbered plans and drawings with a deep discount before the event happens. They had different support tiers ranging from postcard to one of the first ten copies of each plan and have a thriving community of collectors.

Yes: as usual with disease the poorest countries are going to be hit the hardest. If you think that Lombardy's medical facilities are bad, this will be quite the next level.

Is this the hype train? Yes to a large extent. However some people at a16z do know what they are talking about. Ex: Vijay Pande is a partner there and an (adjunct) professor of Bioengineering at Stanford. He and the team around him do have an educated guess how biotech today is different from 20 years ago.

It depends on the type of derivatives. Consider commodity products: basic materials such as iron ore or coal or softs such as cocoa or wheat. Their (spot) prices used to be extremely volatile. Coincidentally, one of the first big economic crisis was caused by a pump and dump scheme in tulip seeds: https://en.wikipedia.org/wiki/Tulip_mania

In this case, futures contracts for those commodity have helped tremendously lower the volatility and make everybody more responsible. It's a good question if the same is going to happen with futures contracts for crypto

Yes. I am OK with accepting tracking cookies on other sites but I just close the tabs if they I see anything related to Oath as a matter of principle.

Every doctor tends to have a very static sensitivity-specificity preferences (true positive rate aka recall and true negative rate, respectively). One of the interesting consequences of using an automated diagnostic tool (already mentioned in Esteva et al 2017's Nature article) is that the sensitivity level can be chosen dynamically, depending on additional risk factors.

That's the whole concept of kaizen, or improvement: https://en.wikipedia.org/wiki/Kaizen

McKinsey has a lot of Chinese walls internally: most partners and consultants who have been working for one client would never advise that client's competitors. The firm is very careful about keeping the trust their clients have in them, as that's ultimately what they're selling.

TLO is a Zerg player, so he probably does a lot more errors when playing Protoss. Also, every top player estimates when to do a sequence of actions and spams it a few times to maximize the chance of execution. Meanwhile Alphastar only has to do that once.

Just a reminder that Google, or any company for that matter, has a fiduciary duty towards their shareholders to file a patent for everything they have learned about and that hasn't been patented yet.

David Freedman has this following dialogue in his Statistical Models: Theory and Practice book:

Philosophers' stones in the early twenty-first century Correlation, partial correlation, cross lagged correlation, principal components, factor analysis, OLS, GLS, PLS, IISLS, IIISLS, IVLS, LIML, SEM, HLM, HMM, GMM, ANOVA, MANOVA, Meta-analysis, logits, probits, ridits, tobits, RESET, DFITS, AIC, BIC, MAXNET, MDL, VAR, AR, ARIMA, ARFIMA, ARCH, GARCH, LISREL[...]...

The modeler's response We know all this. Nothing is perfect. Linearity has to be a good first approximation. Log linearity has to be a good secont approximation. THe assumptions are reasonable. The assumptions don't matter. The assumptions are conservative. You can't prove the assumptions are wrong. The biases will cancel. We can model the biases. We're only doing what everybody else does. Now we use more sophisticated techniques. If we don't do it, someone else will. What would you do? The decision-maker has to be better off with us than without us. We all have mental models. Not using a model is still a mode. The models aren't totally useless. You have to do the best you can with the data. You have to make assumptions in order to make progress. You have to give the models the benefit of the doubt. Where's the harm?

In yoga, this technique is called "virtual" Nadi Shodana or alternative nostril breathing. There are many variants, A/B tested for centuries, that might work better for different people.

Could AMD open-sourcing their (hopefully incoming) stack of learning libraries be a competitive edge for them? They are far behind but this could be an important step.

There is also something else: some funds have become so good in putting the right orders in the book that they don't need statistics so much to generate alpha.

HFT is best defined as predicting movements under a tick; everything above a tick is out of HFT territory. Latency requirements could still be pretty serious though, as many models at many firms can act simultaneously for longer frequencies as well.

Meanwhile Google already has their Quantum AI lab. They're busy with Numerics Flow, a Tensor Flow-like quantum machine learning framework which looks great. Anybody else active in this field?

This is exactly what Questers do. https://www.questers.com/

While that's certainly a concern, let's not forget that Alcantara is very common in high-end sports cars. It's commonly used in endurance racing (24h of le Mans etc).

That's the one :)

This is more of a graduate-level text though. It's as close as you can currently get to Michael Jordan's forthcoming book on graphical models and ML.