tfgg

Blockchain and IoT are so 'B' ark.

AFAIK the health stuff at DeepMind is just one part, the research division is the main bit. Not like IBM Watson consultancy-pretending-to-be-AI-research.

> I'm just not sure what else separates them from any other "AI" (read machine learning/algorithm-building/efficiency consultancy) out there.

probably the amount of pure research they produce: https://deepmind.com/research/publications/

The Casimir effect, contrary to popular understanding, doesn't require reference to vacuum zero point energy, it's a relativistic quantum effect: http://journals.aps.org/prd/abstract/10.1103/PhysRevD.72.021...

No. Quantum field theories and their vacuum state has nothing to do with pilot wave theory (PWT).

PWT is oddly popular amongst internet people, but is just a mathematical transformation of the Schrodinger equation, with some associated interpretation. It makes no additional predictions.

Dark matter is pretty well defined within a particular model, not magic. You can take lambda-CDM, a very simple model, run a simulation, and compare it to observations.

I can't really comment on why MOND is magic. That seems a bit unfair, it's just a phenomenological modification of F=ma, and can similarly be simulated and compared to observations.

Things that make me just close tabs straight away:

1) Autoplaying videos, especially ads.

2) Pop-ups / overlays.

3) Loading lots of extra elements causing text to jump around.

Especially true if I'm just browsing around and click on something that looks interesting, the above will take that thing from "this might be worth 30 seconds" to "not worth it".

Why kill that little dopamine boost someone just got from clicking on a link to your site? If you're wondering why your bounce rate is so high... though maybe these dark patterns bring in enough ad revenue that it's worth it. I don't see how 3 helps that, though, just lazy coding. Or maybe other people are more tolerant and it doesn't really affect the bounce rate.

At https://democracyclub.org.uk/, for the UK, we ran live election results for our open data feeds on our YourNextMP project. That was basically just someone sitting in front of the TV putting in results as they came in :)

If 18F means a single healthcare.gov disaster is averted, doesn't it basically pay for itself?

It sounds like Tegmark is pointing out a pretty obvious and deliberately designed property of LSTMs... the entire point of them is to avoid exponentially decaying / exploding gradients and allow propagation of information over longer time-scales.

> - Discuss each bill in the discussion section

> - Join the discussion boards to discuss the latest events and news

> - Vote on bills (similar to Countable)

Based on my experience of other similar sites in other countries, these features will largely be a ghost town. Even if they aren't a ghost town, they can never be representative samples of the US population.

I'd really encourage you to focus on features that are useful even if you only have one user. It's too easy for this sort of site to try to do everything, successful ones tend to be narrowly focused -- I like your notification approach.

> ... unchecked lawmakers who at any time can run wild and criminalize UK citizens ...

Assuming you're referring to the EU Parliament, the EU has little say in UK criminal law -- as opposed to various cooperation agreements and bits of harmonizing how criminal justice is carried out.

https://fullfact.org/europe/what-would-brexit-mean-criminal-...

> consciousness and insight

Being unmeasurable and undefined, and therefore irrelevant to measures of intelligence, "consciousness" is typically not considered part of AGI [1]. Not sure what you mean by "insight".

[1] http://arxiv.org/abs/0712.3329

A doctor I spoke to about CBT said he thought CBT was effective for the relatively intelligent, introspective people they might have had in early trials. Rolling it out further beyond that group had lower success rates.

The additional the article raises about lower standards of training as it scales is also an interesting one, and could possibly apply further to other fashionable "therapies" in things such as fitness and project management.

As someone who was in condensed matter physics, I think this is one of the most important articles for understanding the modern mindset in physics, and even fields like machine learning.

Seeing how something can be based on the rules of something else, but have novel emergent behaviour, opens up a whole new ways of thinking about phenomena.

It certainly always elicits a lot of "Hurdur this is my company", feels a bit too perfectly written for that reaction to not be satire.

Weird coincidence, but I remember someone at Cambridge (Christian Steinruecken, http://www.siski.de/~chris/) did his undergraduate dissertation on a Quantum ML. Sadly all the links I can find to it are dead now :(

Have you considered using a TrueSkill-alike with extensions for scores/teams, such as PoissonOD?

http://research.microsoft.com/pubs/193839/sbsl_ecml2012.pdf

A few years ago I had a similar idea for trying to build team models, so you can make a better guess at the performance of national teams, since they don't play very often, or for league teams due to transfers at the start/during the season, but hadn't got as far as you :)

How does this square with MACHO (non-)observations? [1]

> A MACHO may be detected when it passes in front of or nearly in front of a star and the MACHO's gravity bends the light, causing the star to appear brighter in an example of gravitational lensing known as gravitational microlensing. Several groups have searched for MACHOs by searching for the microlensing amplification of light. These groups have ruled out dark matter being explained by MACHOs with mass in the range 1×10−8 solar masses (0.3 lunar masses) to 100 solar masses.

[1] https://en.wikipedia.org/wiki/Massive_compact_halo_object

It's the number represented by that file's full binary value (all the bytes concatenated together).

Right, and as far as we know, quantum mechanics evolves the wave function in a time-reversible fashion.

Also one of the intelligences encountered in Diaspora by Greg Egan.

Actually, they just meet a localised handling agent that explains that they work for a star-spanning sentience - which is something the characters had been considering as a way of making themselves robust against civilisation-destroying gamma ray bursts, but dismissed on the grounds that it made everything so slow - though that links into the other theme in that book of simulated beings existing on very different timescales.

Lovecraft does data science?

(the variance on the graphs intuitively felt a bit low, but then again it is facebook)

I feel comparisons to the Apollo and Manhatten projects, often brought up in this context, are misleading because they ignore the massive pure research bases they were building on. Of course they did do more pure research, but a lot of the cost and muscle/size was for solving engineering problems.

Sorry, I think this works: https://lagunita.stanford.edu/courses/HumanitiesSciences/Sta...

Good points. I'm coming from a computational physics research background and applied for a few data science positions at startups, a large social network and a private ML research group, so not that many overall, beware of the small sample size.

The smaller startups seemed to want more "data engineering" experience.

What do you mean by "what if any bias was there in your job search"?

If you have time, I'd also read something like David Mackay's information theory textbook [1] for more of a Bayesian perspective. Interviewers did seem to appreciate having multiple perspectives and interpretations of basic results, though less practical.

[1] http://www.inference.phy.cam.ac.uk/mackay/itila/

Based on recent (successful) job interviewing, I'd recommend people looking for a job in data science/ML to do a statistical learning course such as the Hastie and Tibshirani Stanford one [1] as a higher priority over ML/deep learning courses. It gives you a base level of knowledge in the field, and even for jobs that do deep learning, most of the technical questions will be about making sure you know the classical concepts really well.

[1] https://lagunita.stanford.edu/courses/HumanitiesSciences/Sta...

You don't know his baseline personality/attitude. It might be he's good at the relaxing and being creative aspect, and so needs to personally remind himself to work hard. Other people might naturally work very hard, but need to remind themselves to relax.

Huh? Orthodontics are free in the UK up til 19, and it's very normal to have braces.

Maybe amongst an older generation who'd have to go out their way to pay for straightening.

In general, minimising a problem with the universal electron-correlation functional is known to in NP due to this [1]. Which is kind of sad for density functional theory (DFT). I've probably misphrased that.

Also, side remark: Coding up a 2D Ising model solver with Metropolis-Hastings is surprisingly fun. Doubly so if you GPU accelerate it.

[1] http://www.nature.com/nphys/journal/v5/n10/abs/nphys1370.htm...