Disruptive Proactivity

So it’s been fifteen glorious years, and while we may all have aged (or been born, for some regular attendees), and the political system has been ‘challenged’, and atmospheric CO2 may have gone from 399.77 to 411.75ppm, we persist, and we plan to head off up river again on Sunday 30th June 2019.

The first Geek Punt Picnic took place in 2003, and in 2004 it morphed into Puntcon, and was announced in an email in June 2004 as follows:

NotCon was fun, but didn’t have water. Or champagne. So this year’s first Geek Punt Picnic will also coincide with PuntCon ’04, the Cambridge leg of the alternative conference circuit.  In a small break with tradition there will be no talks, no presentations, no agenda and nothing to disturb the quiet delights of the river on a Sunday afternoon. But apart from that, it’s a conference and therefore probably tax-deductible.

We’d love to see you on the 30th June 2019.

The Official Invitation:

You are invited to Puntcon, the fifteenth great geek punt picnic, taking place on or about the River Cam on the afternoon of Sunday 30th June 2019.

Full details are here on Eventbrite. There’s a Facebook event, too, for those who want it.

Although it says there are 250 tickets that is ‘nominal’ as it’s a big river… but Eventbrite needed a number 🙂

How it Works:

As before, turn up outside the Mill public house on Mill Lane between 1200 and 1230. (https://map.what3words.com/deck.jungle.junior)

Four or five people will be deputised as PUNTERS and will come with Sam and Bill to collect the boats, the others will walk up beyond the rollers where we board https://map.what3words.com/vibrate.galaxy.volume(See this useful post about What3Words from Terence)

We will head off  between 1230 and 1300 – if you are late you can walk up river and catch us as we don’t punt very fast!  We will be heading upriver rather than along the Backs – more picnic places, fewer tourists. We normally stop half-way along Grantchester Meadows and hum Pink Floyd tunes.

Bring something to drink and something to eat. Actually, bring a lot to eat and a lot to drink – it’s thirsty work!

Note that there aren’t any toilets (but there are hedges) – the nearest pub/tea room is approx 20 min walk)

Punts will be booked/organised/commandeered for you so you do not need to hire one. We will take as many punts as we need – we get four or five of the big punt that hold 12 and then extra six-person ones –  and head up river to a convenient picnic place on Grantchester Meadows where we will eat/drink/carouse. Those arriving late can join us there.

Anyone can punt. Anyone can be shown how to punt. But you are not expected to punt. While we support the consumption of apples in all forms, please keep electronic devices in the punt and not in the river.

We normally get back to Scudamores around 1800. Those heading back to the railway station can be dropped at a bridge within walking distance.

Post punting we have the option of retiring to the pub or a restaurant and letting Sunday evening happen around us.

How Much

You are asked to contribute your share of the punt hire. There is no registration fee. Bring food and drink and entertainment. Punt fare will be split between all comers –  it works out at twenty pounds per person. Infants are not expected to contribute.

The Invitees

Please feel free to invite other people – it’s a big river and there are lots of punts. This page is here to give us a rough idea of numbers in advance

Some idea of who’s coming is available via The Facebook (unfortunately, other platforms aren’t really available, yet). Book your spot in a punt via eventbrite.

So come along — it’ll be fun.

Cheers

@billt and @smithsam

Hal writes:

Hassabis often cites Breakout, a videogame for the Atari console. A Breakout player controls a bat that she can move horizontally across the bottom of the screen, using it to bounce a ball against blocks that hover above it, destroying them on impact. The player wins when all blocks are obliterated. She loses if she misses the ball with the bat. Without human instruction, DeepMind’s program not only learned to play the game but also worked out how to cannon the ball into the space behind the blocks, taking advantage of rebounds to break more blocks. This, Hassabis says, demonstrates the power of reinforcement learning and  the preternatural ability of DeepMind’s computer programs.

It’s an impressive demo. But Hassabis leaves a few things out. If the virtual paddle were moved even fractionally higher, the program would fail. The skill learned by DeepMind’s program is so restricted that it cannot react even to tiny changes to the environment that a person would take in their stride – at least not without thousands more rounds of reinforcement learning. But the world has jitter like this built into it.

Hassabis leave more than a few things out, as Dr Beth Singler also points out.

Anyone who played breakout enough in the early 90s found that the thing that demonstrates ‘the preternatural ability of DeepMind’ was possible and effective at winning. It may not be immediately obvious to those who haven’t played the game, and who don’t understand that this is a natural result of playing games with scores. Whether it was new to engineers in 2015 is unclear, but it was new enough to the DeepMind PR people that they persuaded journalists it was actually novel – did anyone who said so ever play the original game?

This novelty bias also applies to the AlphaGo move that DeepMind PR pushed in 2016 – that ‘unique move’ appears a bunch of times in the KGS archive.

As Whitehall Departments think they should sprinkle AI in their spending round submissions, there is another lesson in Hal’s words.

When we talk about expensive deepmind generalists over cheap NHS specialists, this is what we man. The civil service also believes expensive generalists over their own specialists too. In both cases, the expensive generalists assume what is new to them is new to everyone. In medicine, the specialists know better, but it’s harder to write fawning articles about that.