Sunday, November 12, 2017

Not Stupid - Scary

This is not a windsurfing post. Contrary to what some of you may have thought after reading the title, it's not a political post, either. This post is about something that scares me - and scares me quite a bit, more than the front loop in windsurfing.

I develop software for a living, and a lot of my work deals with complex algorithms. At times, that includes machine learning and artificial intelligence methods like neural networks. A few years ago, it seemed that computers were stupid and would remain stupid.

That has changed. Now:
  • Computers can act with more intelligence than any human
  • Computers can learn such intelligence-related skills hundred of times faster than humans, without any supervision - computers can learn skills that take humans decades to master in weeks.
These statements are based on recent developments in computers playing games - specifically, the game of Go. Go is a board game that is very popular in East Asia, where it is played by more than 40 million people. This includes more than one thousand professional Go players in Japan, China, North Korea, and Taiwan. They compete in a number of tournaments where the winner's purse can be as high as $500,000 (compared to the total prize money of $140,000 at the largest windsurfing event, the PWA World Cup in Sylt).

Compare to chess, Go has much simpler rules. But while computers have been able to beat chess champions since the 1990s, Go has been a much harder problem, partly do to the large number of possible moves that make a "brute force" approach to finding the best move impossible. It took until October 2015 until a computer beat a professional go player in an even match.  A few month later, the next version of the computer program beat an 18-times Go world champion.

This was an impressive feat, but the story does not end here - it gets better. The version of the software was able to run on a single computer rather than a network of computers that previous versions required; it beat professional go players 60:0.

Then came the really crazy improvement: AlphaGo Zero. Whereas pevious versions had been trained with thousands of Go-games played by amateur and professional players, AlphaGo Zero only knew the rules. Over a few days, it played a few million games against itself, and used the outcome of the games for "unsupervised learning". After 40 days, AlphaGo Zero played against the older version that had beaten the world champion.  AlphaGo Zero won 100 out of 100 games!

So - a computer program taught itself in 40 days to reach a level that takes the best human players decades to achieve! That's absolutely amazing.

It's also very scary. If a computer can teach itself to surpass any human at a very difficult mental task within weeks, then "artificial intelligences" that are generally more intelligent than humans suddenly don't look like science fiction anymore. Some of the most intelligent people on this planet, including Steven Hawking and Elon Musk, have warned about the potential dangers - perhaps it would make sense trying to understand what they are concerned about?

I won't delve into that now, but let me give you a few things to think about. The AlphaGo software was developed by Google, and is running on hardware designed by Google. One computer with 4 "TPUs" can beat the best human Go player; in total, Google uses about 2.5 million servers at it's gigantic data centers. Plenty of computing power to learn other things. How about learning about the ethics of one species exterminating tens of thousands of species?

Of course, we don't really have to worry about computers - they can't harm us because we can just turn them off, right? Only if the computers were somehow connected to weapons would be have to worry about those science-fiction scenarios. There may be some military drones around, but they are usually flown by human operators; even if capable of autonomous flight, any firing decisions usually require a human. According to Wikipedia, the "U.S. Military is investing heavily in research and development towards testing and deploying increasingly automated systems". But thankfully, these are still to be controlled by human beings, thoughtfully supervised by the Commander in Chief. Nothing to worry about!
If you're interested in learning more about the underlying AI or want to watch the Go games between AlphaGo and Go professionals, check out the DeepMind channel on YouTube.

Friday, November 10, 2017

Fun in Lüderitz

The speedsurfing event at the channel in Lüderitz got a lot better this year - Ben Proffitt is there, and he keeps up up-to-date with daily videos on  Today was a light-wind day, so he gives a nice overview of the channel - with short segments of impressive crashes from previous years to illustrate the problems at the different areas. Enjoy!

Wednesday, November 1, 2017

Twenty, Forty, Thirty

Three days of wind in a row - very different days.

20 mph ESE at West Dennis. After 2 1/2 weeks in Hatteras, it feels very bumpy. The meter seems to read high - I need my 117 l board to get planing with a 7.0. Nina is barely powered on her 5.0, and "everything felt wrong". Good that every day of sailing is a great day of sailing! In the evening, we drive to Mystic, CT, so sail the Sandy Point Slicks on Monday.

A huge storm brought 70+ mph winds during the night, with power outages from North Carolina to Canada.
Getting to the Barn Island boat launch takes longer than planned since some streets are closed due to fallen trees.
We rig when the meter readings show 32 mph averages, and some computer models predict further drops. Nina rigs the 3.4, our smallest sail. I rig the 5.6 Racing Blade, since Boro said the sail can handle a lot of wind. Sailing away from the launch is easy - there's a pronounced wind shadow. Maybe it should have tipped me off that I was fully planing within a second on the 72 l speed board. Once I leave the wind shadow, the fight starts. The sail is too big! Crash! Try to waterstart. The sail pushes me under water. I thought I knew how to waterstart? Again. The wind picks up the board and throws it around. And so it continues. Sailing out 2000 feet took 90 seconds. Getting back takes half an hour.

I rig down to the 5.0 Koncept, and fight my way up to the sandbar through 1-2 ft chop. On my way, I set a record - for the slowest sailing of a speedboard ever! But even at just 10-15 knots, it only takes 15 minutes to get half a mile upwind and a mile across. But when I arrive, I need a break! I'm also disoriented, and spend the next 40 minutes walking around looking for flat water. I later discover that the wind had picked up to averages above 40 mph, gusting above 50, during that time. Nina spent a lot of this time sitting on shore - the 3.4 is way to big!

And so it continues. We start sailing, the wind drops. I'm nervous about sailing back a mile on a 72 l board - I need about 110 l to float me! I sail back, rig the 5.6 again, and pull out a big board. As I get ready to go out, the wind picks up again to 40 mph. So back to the 72 l board, and once again an upwind sail to the sand bar. I get a few ok runs in before the wind drops. While I was going back and forth, Nina had some fun working on Vulcans on the 3.4; Dean set a new personal best for 2 second top speed of 37.7 knots; and Bart, who arrived late due to many closed streets, hit 35.8 knots. By the time I get things dialed in, the wind starts dropping again, and I barely manage a 32-knot run. But at least I managed to make it back to the launch without having to slog. Every day of sailing is a great day of sailing, but this day was an adventure.

The wind is finally nice. It lets us sleep late, and then comes in exactly as predicted (after adding a few miles per hour to the forecast because it's WSW). After feeling a bit like a beginner the day before, I decide to show the 5.6 who's the boss! I ask Nina for rigging advice. With almost an inch more downhaul, the sail actually has some loose leech! I still bounce around a bit when crossing over to Egg Island, which is half a mile downwind:

But once I get there, it's flat! The wind is almost at a right angle to the second sandbar, which is fully out due to a very low tide. That's not great for top speed, since going deep downwind means hitting chop after just a few seconds; but it's fun for just going back and forth, and working on jibes. I'm having a blast - compared to the 72 l speed board, the 90 l slalom board feels huge, and is really easy to get going and jibe. I end up with my 3rd-fastest session ever (33.5 knots, 62 km/h), and tie my personal best for alpha 500 with 22.5 knots. Here's a video of this alpha 500 run:
Average speed on the first leg was 29.6 knots; minimum speed in the jibe was 10.7 knots; and average speed on the second leg was about 19 knots. There's lots of room for improvement in the jibe and coming out of the jibe ... next time!

I managed to sail back in time just before the wind dropped too low, and had too walk and slog just a little bit - well worth it! That was a great day, all alone at Egg Island.

Sunday, October 15, 2017

Fighting Chicken

Our first five days in Hatteras were windy - we planed every day, usually in wind that was higher than the forecast. Friday took the crown - the forecast predicted 20 mph at 8 going down to 15 during the day, but we got 23 mph going up to 30 and staying there.

I took the opportunity to fight my inner chicken. It's quite a loud-mouthed beast, and I often listen to it. Not this time! Out came the 72 l speed board, even though the wind had turned from NNE to NNW by noon, meaning higher water levels and more chop. I had been overpowered towards the end of the morning freestyle session on my 5.6 m sail in 25 mph wind. With meter readings of 29 mph after lunch, switching to a 6.3 m race sail sounded about right. You need to go big for speed, right?

Well, the 6.3 turned out to be encouraging the chicken just a bit too much. In theory, a 6.3 should be easy to hold in this wind and "moderate" chop. Indeed, the sail never felt overpowering; instead, it combined the best characteristics of a freestyle sail with amazing stability. But with the inner chicken constantly screaming "Not so fast!", I never quite got the feeling of being in control.

On the verge of once again loosing the chicken fight, I rigged down to a 5.6 Loft racing sail. That made a lot of difference, and I was finally able to get a few runs near 30 knots. That may be nothing to good speedsurfers, but for me, it's a lot in chop. With the chicken finally under control, I even found a patch of flat-ish water near shore, and get a 32 knot run - 2 knots faster than my previous best in Hatteras, and my 5th-fastest session ever. Take that, chicken!

We'll be at the ABK camp for the next five days, so we'll focus on freestyle instead of speed. Maybe I can beat the chicken once again and try a few loops...