469 karma · joined April 12, 2009
http://www.github.com/batterseapower http://blog.omega-prime.co.uk/
After Epic's wild success hes now got a bit of cash to splash on making his dream a reality :-)
However, if I then simply ask "What is the most probable result for this function to return?" it figures out the answer and a very good approximation of the probability (4.5e-5). From there it's easily able to rewrite the program to use the trick. So the creative step of spotting that this line of reasoning might be profitable seems missing for now, but 2025's models might solve this :-)
The curmudgeon in my questions where it really make sense for the Caltrain to exist at all, given that it:
1. It is extremely far from covering its own costs
2. The subsidy provided by the state is regressive because it principally benefits the wealthy commuters along the line
3. It imposes substantial negative externialities on the South Bay both in terms of noise & traffic delays at grade crossings
4. The line and stations occupy very valuable real estate that could be sold and put to better usesThe other idea I had when I was working on this stuff was to do JIT supercompliation. Clasically, supercompilers expand the whole graph at compile time, but it would be straightforward to delay the supercompilation process until the first time that the code is actually executed, which helps tame the problem of generating lots of specializations that may never actually be executed in practice.
(Choosing a generalization heuristic becomes more important in the JIT setting because you always have access to the full info about all parameters to the function you are specializing, and naively you would end up e.g. specializing sigmoid(x)=1/(1+exp(x)) for each value of x observed at runtime.)
Cool to see that people are still working on it, but I think that the main barrier to practical use of these techniques still remains unsolved. The problem is that it's just so easy for the supercomplier to go into some crazy exponential blowup of function unfolding, making the compilation step take impractically long.
Even if you avoid a literal exponential blowup you can easily end up generating tons of specializations that bloat your code cache but don't reveal any useful optimization opportunities/are infrequently used in practice. Similar performance problems also plague related techniques like trace-based JITs, even though the trace JIT happens at runtime and thus has access to strictly more information about the frequency with which a trace might be used.
You can try to use annotations like the @extract proposed in the article to control these problems, but it can be hard to predict in advance when this is going to occur.
One interesting research direction might be to use deep reinforcement learning to try to guide the generalization/termination decisions, where the reward is based on A) whether the unfolding leads to a tie-back later on, and B) to what extent the unfolding allowed deforestation/beta reduction to take place.
A recent properly-conducted randomized controlled trial found that WFH improves productivity: https://www.nber.org/papers/w18871
The relevant keyword if you want to learn more here is "equirecursive types".
The solar flare risk (at ~1% per year [1]) seems much more likely than many of the other items there. Supervolcanos, pandemics, war, and AI also seem like salient risks, but we probably shouldn't worry about e.g. gamma ray bursts or asteroids.
[1] https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/201...
[1] https://toddwschneider.com/dashboards/nyc-taxi-ridehailing-u...
The basic idea is to use word vector embeddings to build a source<->target dictionary, then combine this with a language recognition model to iteratively bootstrap a set of source<->target training examples for use with a conventional ML approach.
- Wage inequality (the average wage is growing faster than the median wage)
- Growth in benefits like pensions and health insurance which are not part of wages but are part of compensation
The inequality and benefits effects are about equally strong. Once you control for these 2 factors, compensation has risen at almost exactly the same rate as productivity growth, which is what you would expect of the economy in long run equilibrium.
[1] https://www.resolutionfoundation.org/app/uploads/2014/08/Dec...
Meta-analysis suggests that supplementation in vitamin-D deficient populations does not reduce cancer, fractures, cardiovascular problems or all cause mortality (http://www.sciencedirect.com/science/article/pii/S2213858713..., https://academic.oup.com/jcem/article/96/7/1931/2833735).
One limitation of this research is that most studies are done with fairly low doses (700-800 IU is common). However, those few studies that look at higher doses (5000-1000IUs) you don't necessarily see any better signs. On the contrary, there is some evidence that high doses seem to be associated with e.g. higher risk of falls and fractures in elderly patient populations (https://jamanetwork.com/journals/jama/fullarticle/185854, https://jamanetwork.com/journals/jamainternalmedicine/fullar...).