While Apple are known for custom cooling designs they are also known for building their notebooks very close to the limit when it comes to dissipating heat. So when nVidia twists the truth about the operating temperature of a prospective GPU for their macbook it can have bad consequences. This combined with the well known faults caused by the bad solder bumps on nVidia GPU die has caused the relatively large proportion of early GPU failure rates in various macbook models. Apple and nVidia have tried to mitigate this by effectively under clocking various models through a firmware update to delay the inevitable failure.
Being an owner of such a model I can tell you it's pretty easy to get the fans whiring when using the more powerful GPU and that's even after the under clocking firmware update. Considering how common this is, i'd say it's likely that the heat pipe is fine, the failure rate is accelerated enough with the other two problems..
"SMTP has been designed so that it has no single point of failure"
From SMTP:
"An e-mail client needs to know the IP address of its initial SMTP server and this has to be given as part of its configuration (usually given as a DNS name). This server will deliver outgoing messages on behalf of the user."
...how?
The internet needs a big threatening stick to cut the legacy cord of DNS and start the evolution to a decentralised system.
There are enough existing protocols to draw upon that contain conceptual components of a distributed DNS: Zero config DNS, Bittorrent etc.
This made me think... everyone always discusses the rules and laws for drones and their co-existence with larger aircraft... which is fine. However that's not going to stop anyone who doesn't mind breaking the rules (intentionally or otherwise), drones are relatively cheap, no license is required to get one, and there's pretty much nothing physically enforcing the use of drones.
So what about enforcing those rules when it comes to the larger non military manned aircraft... Perhaps they should have some kind of basic defence against small unmanned craft in their flight path. I wonder what kind of "airline" level weaponry would be acceptable to take out drones while not posing a risk to military aircraft. For starters an on-board radar and tracking system would be needed to pick up drones that are too small to be visible to ground based radar... It could even make the occasional encounter with bird shaped projectiles a little less dangerous.
The only problem with shooting a drone out of the sky is of course falling bits of drone... so i guess avoidance would be preferable given a sophisticated enough on board radar.
"The virus is slowing down in its ability to cause disease and that will help contribute to elimination."
Not sure if this is incorrectly phrased or i'm incorrect in my understanding of viral evolution... The virus has evolved to slow down the process of causing disease, surely this is because it is advantageous to the continuation of this virus, if the host dies too quickly they are less likely to pass on the virus. So how does this contribute to eliminating the virus? is it not the opposite? Longer infected lifespan == Greater chance of transmission.
Adding another layer of material will decrease transmission of light in the opposite way that surface patterns work to increase the transmission of light: Any medium with a refractive index different to air will reflect some percentage of incident rays that are not perpendicular to the surface, surface patterns can help avoid this by effectively changing the angle of incidence to be closer to the surface normal or redirecting some of the reflected rays.
It would be better if the surface doubled as the protective layer. Provided the pattern is small enough or better yet if a pattern was chosen which is also poor at allowing particular sized molecules (i.e water) to settle on it's surface in the same way that hydrophobic surfaces work... then it should be perfectly suitable as a protective layer also.
Let's say, however, you built a virtual world at a reasonably fine-grain (doesn't have to be too fine, just good enough), a second virtual world that was much coarser-grain and which used lossy encoding in a way that preserved some information from all prior states, a crude set of genetic algorithms that mapped outer virtual world to inner virtual world, and finally an independent set of genetic algorithms that decide what to do (but not how), a set for examining the internal virtual world for past examples of how, a set for generating an alternative method for how without recourse for memory, and a final set for picking the method that sounds best and implementing it, and an extensive set that initially starts off with reconciling differences between what was expected and what happened.
That should be sufficient for Emergent Intelligence of some sort to evolve.
Perhaps, but there is still quite a lot of pre-defined structure there. Although i've no doubt that some pre-defined structure is far more pragmatic and likely to yield useful results than what i'm thinking of (and i have given this some thought previously). It's difficult to know what an environment with emergent properties suited to a digital medium should look like, because it's so different from the vastly more complex environment that biology emerged from. Which could perhaps be summed up in three parts:
Finding a reasonable equivalent to spacial dimensions seem simple enough, and probability is already the basis of most genetic algorithms, but the building blocks... the rules of resources and state are massive and complex as chemical reactions. I think finding a simpler mathematical equivalent to those structures that also has the necessary emergent properties is fundamental to creating an emergent AI.
It's not enough to emulate the properties of intelligence, you have to emulate the reason for there needing to be intelligence in the first place.
This difference was clear to me when reading up on existing AI and machine learning methods.
AI in it's current form feels more like engineering than an exploration in nature, science and math. Slightly dangerously with my limited knowledge in AI i would describe AI today as an extremely useful and insightful set of tools inspired by nature, but which are not themselves nature. They are just yet another thing that we have learnt to re-implement as a fruit of biology. Actually cellular automata feel more like nature than AI.
Methods such as neural networks are pre-evolved static solutions, the information flowing through them may evolve, but the method which determines their flow does not itself evolve, they are therefore selective and static imitations of the a brain much the same as an animatronic manikin is an imitation of the body at an evolutionary static point in time.
It's conceivable that with enough detailed imitation an intelligent implementation of a whole brain (not even human) could be achieved... but it seems highly unlikely and impractical. However implementing the basis or conditions to give emergent and undirected development in a "synthetic" medium would be nature at work or "life" in my view. Imagine an AI that had the freedom and incentive to create it's own methods dynamically, that kind of creative freedom must be a pretty good fit for true "Artificial Life", so shouldn't it be called "Emergent Intelligence"... The opposite to "Imitated Intelligence".
Tony and Jan Jenkinson have not been told whether they will get the £100 charge refunded, following the withdrawal of the charge
Good that the trade regulators stepped in however it seems like the family are still owed their £100. The point is not that it's a massive amount of money, more that they should never have charged them at all, so they should give it back.
The one day you'd sell your soul for something, souls are a glut.
