Me, on DeepMind plans for rat-level AI, and DeepMind is making breakthroughs at an alarming rate.

Only 80%?

Still better than allowing diesel cancer to spread wildly in the population... that's going to be DDT of 21st century: looked miraculous at the first look, turned deadly Satan's invention. In the article, risks are admitted, but severity dismissed (hard to prove wrt.. DDT)

How do diesel exhaust fumes cause cancer?

When diesel burns inside an engine it releases two potentially cancer-causing things: microscopic soot particles, and chemicals called ‘polycyclic aromatic hydrocarbons’, or PAHs. According to Phillips, there are three possible ways these can cause cancer:

“Firstly, inhaled PAHs could directly damage the DNA in the cells of our lungs – leading to cancer.

“Secondly, the soot particles can get lodged deep inside the lungs, causing long-term inflammation, and thirdly this can increase the rate at which cells divide. So if any nearby lung cells pick up random mutations, this inflammation could, theoretically, make them more likely to grow and spread.

Read more at http://scienceblog.cancerresearchuk.org/2012/06/14/diesel-fumes-definitely-cause-cancer-should-we-be-worried/#BspXAGUiOMAgtljS.99

Well, we humanimals should realize that our last biological stage in the Glorious Screenplay of the Evolution of Intelligence is here solely for the purpose of creating our (non-bio) successor before we manage to destroy ourselves (and probably the whole Earth) - since the power those humanimals get (especially the ruling ones) from the memetic supercivilization living on top of the humanimalistic noise (and is represented by a fraction of per cent that gives these humanimals essentially for free all these great ideas, science and subsequent inventions and gadgets) is rapidly becoming so huge (obviously, exponentially) that we have simply no chance to manage the upcoming nano revolution... already nuclear bombs were just-just - now imagine affordable one dollar DYI nuclear grenade every teenager can put together in the garage... nanobots will be orders of magnitude worse.

So, maybe we are first on Earth, but not necessarily last... if the Singularity does not make it. We are the threat, not "rogue AI".

Please, keep this secret and do not tell the ruling politico-oligarchical predators... or else you will see how "creatively" can our beloved financial sharks play with it... start with brutally leveraged subprime asteroid extinction risk contracts for 1000 years (plus one can easily imagine lobbyists forcing government to stop funding anti-asteroid research/tech so that it does not harm their business (otherwise, we will hear these spells again... THE WHOLE global financial system will go down and the whole economy with it)

Already one of the best pragmatic statisticians, J. V. Stalin, knew that "One dead is tragedy, million dead is statistics".

Just a little remark: before "friendly AI" becomes a problem, there's the usual one: "friendly humans".

Let's take the standard assumption: 1. Evolution is about evolution of Intelligence. Needs our last animal/wetware stage only to develop its "in silico" successor.

Then the conclusion is quite obvious: 2. We are the danger. Mostly because due to our stupidity we can easily destroy ourselves before advancing to the next level.

It's quite a simplecstory: animals got technology from another, higher civilization - this smells deadly.

Memetic parallel civilization of a tiny intelligent but powerless fraction of per cent is giving the humanimal "society" (for free and without any say/power over its usage) more and more advanced power based on sci/tech.

Enormous conflict coming from the discrepancy between these advances and animalistic roots of the "society" (ruled by the usual predator/prey jungle-system - call it king/slaves, employer/employees or politician/taxpayers) is clear.

Already nuclear bombs were beyond the limit. Now imagine cheap, almost homemade drone-transported nuclear grenade - coming nanotech can be much worse, as we all know.

Now slip just for a while into the boots of the upcoming Super AGI.

Giving this techpower to the bunch of predatory lunatics (read: governments) often drugged by stupid ideologies? Or even general public?

Are you kidding?

We are much bigger existential risk for the Superintelligence than the AGI is for us (and if it's really THE Superintelligence, then the humans do not matter anymore, anyway).

I'm afraid we will never know whether someone is "close" to (super)human AGI, unless this entity reveals it. Now think nuclear bomb... and superAGI is supposed to be orders of magnitude more powerful/dangerous.

So, not unlike the wartime disappearance of scientific articles on nuclear topics, certain (sudden?) lack of progress reporting press could be an indicator.