Indeed, you are right. The DSM-IV is still in use but, technically speaking, it has been replaced by the newer DSM V, which also includes a diagnosis of caffeine intoxication. I don't have a DSM V handbook within reach, but I'm sure you can find out the details.

Does the fact that people suffer from side effects before dying surprise you? Using the reference you provided above, people die (LD50) at about 490g of pure alcohol (ie >1lt of hard liquor), yet most would agree that toxicity is apparent way before that amount. Caffeine is similar.

[quote]
Toxic levels of caffeine: 12,000 mg (for an 176 pound person) [just-think-it.com].
[/quote]
You are right with respect to the LD50, obviously referring to lethality. However, 1000mg of caffeine is certainly sufficient for caffeine intoxication as per the DSM-IV disorder (code 305.90). It all depends on how you interpret "toxic" in this context.

I think there is a subtle difference between being right (in the usual sense of providing a model that happens to accurately represent measurable stuff) and the process of scientific discussion. Consensus is just an outcome of a process, ie collaboration. That process is extremely important but does not guarantee being right.

In the end, without resorting to unnecessary complicated terms, if a bunch of people who are supposed to know what they are saying all agree on something that is not immediately testable (say, long-term human impact on the climate), odds are they are more likely to be right than some random wacko or idiot reporter because they spent some time discussing together and have highlighted potential errors.

In the absence of definitive hard data, which will only be available in retrospect, we have to pick sides. Consensus seems a safer bet than the probability that some random guy is the new Galileo or Einstein.

I imagine taste depends a lot on environmental factors, like soil nutrients, sun and stuff like that. Creating a genetically engineered plant is probably harder than simply improving culture conditions to get a good product. Although you could probably improve some aspects of the product by bio-engineering, in real life I don't think anyone will care to improve taste. Most probably they would sell you cheaper beans that resist infection or transportation or bad climate. I never heard Monsanto advertise their products as "great taste", but then again I wouldn't know.

Adding caffeine is also so simple that you shouldn't have to modify plants to get it. In fact, caffeine is dirt cheap (http://www.alibaba.com/showroom/caffeine.html). Even pharma-grade caffeine for the lab is like $0.50 for an almost toxic dose of 1000mg, if you want to "enrich" your coffee.

Decaf, if you enjoy it, would be an interesting bio-engineering project. I don't know if the plant really needs the caffeine for something else (it is somewhat related to DNA, being itself a purine), but simply deleting or attenuating the gene that catalyzes the conversion should be really simple. In fact, you don't need the whole genome sequence to do that, only the locus of the gene.

Anyway, bio-engineering something that tastes good and is healthy is probably at least as hard as all other aspects of the production, even if you have the DNA sequence. Being a fan of the KISS principle, I'd rather have my coffee prepared by people with some decent traditional know-how.

AMD TrueAudio on Kaveri processors and newer GPUs supposedly does just that. I haven't seen any game supporting it, though. Would be a nice feature I think.

Thanks for the information!

I have tried some modest consumer-level equipment with disappointing results so I was thinking of either going high-end consumer or even pro. In the end, I'm probably going to take the plunge. As you say, I don't want to change equipment every year...

This is an expensive solution, but I am tempted. Is it better than the equivalent top-end consumer grade products like the Netgear R7000 or the Asus RT68? Specifically, I was thinking of the combination Ubiquiti EdgeRouter PoE + Ubiquiti UniFi AP AC which is almost $700 of gear. Is it worth it for a gigabit home network with a 300MBps fiber connection?

Any other experiences with the Nighthawk AC1900? The hardware seems quite good. I don't mind the price, up to about $250, but I plan on getting 300MBps fiber so I need a router that can reliably route that plus handle my gigabit home network.

A classical article on the subject, quite old now, has concluded that approximately 80% of diagnoses can be made from the history (ie a structured interview) with a further maybe 10% from physical examination and maybe 5% from additional investigations (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1673456/). Obviously, the requirements for modern medicine and the available means are a bit different. Nevertheless, any serious doctor will tell you that the history taking and the physical examination are the most important parts of an encounter with a patient.

This is a direct result of Bayes theorem: the interview and physical define the "prior" probability for any diagnosis and any further investigations will only serve to modify it by a certain degree (confirm or exclude). With the exception of some quite aggressive diagnostic methods, like a biopsy or laparoscopy, which will never be recommended upfront, most investigations are generally not sufficiently sensitive or specific enough to give a conclusive diagnosis.

Finally, the physical (which cannot be done via Skype) is also a very important component of the physician and patient relationship. An encounter without physical examination seems, in my humble opinion, quite superficial. Patients are generally more satisfied if you take the time to carefully examine them.

That being said, Skype can be a decent solution for people living in remote areas where transportation can be a real problem. Skype for people living in major cities is a bit silly, I think.

PS. I am physician, but I am curious to hear what you think about the value of physical examination.

Well, if "not doing anything" is what your PC does a long part of the day, idle power consumption can be of some importance. That does not necessarily redeem AMD cpus, but it is worth mentioning in my opinion. Obviously, you seem to think that all discussion should be limited to "AMD sucks". Even if true, this does not make for a very interesting read.

Well, in idle, which is what most processors do in typical user workloads, the 7800 is comparable to intel processors. Total energy to accomplish a task obviously varies, but the 7800 uses 30-50% more energy than intel processors for the same task. However, the 7 series APUs are clearly more efficient than the 6 series Richland APUs that they replace. Peak power consumption is around 100W for a complete system with 7800, which is not a huge thermal load.

In the end, what I'm saying is that AMD improved power efficiency way more than absolute performance, something that is more important in the portable space. You're looking at maybe 10% faster than Richland at the CPU side but with 20% less energy. Obviously, they could have chosen different power/performance tradeoffs, if they wanted to compete on the desktop.

For some numbers, you can have a look at Techreport or Anandtech (http://techreport.com/review/26845/amd-a10-7800-processor-reviewed/3). I am not aware of any tests concerning the laptop variants, but they should appear soon.

Benchmarks from the smaller 7600 only show a modest performance hit from going to 45W, approximately 10%. Isis Pro is a brute force solution to the problem (huge on-chip RAM) and is likely to stay quite expensive for a while because of die-size and limited production. A discrete GPU is probably the better option at that price point (ie replace the cheapest $450 Iris pro with $450 cpu + GPU), unless if low power consumption is an absolute priority.

Affordable on-chip graphics RAM may become standard in future AMD and intel processors but I wouldn't hold my breath...

They will probably make some version for the server market, but it will certainly be on another socket. The socket AM3+ does support ECC (if you choose the right motherboard, ASUS usually do...) but the upgrade path is probably stuck forever at the FX8350. It isn't a bad chip, actually quite good for multithreaded loads, but it's getting old... If you want ECC for cheap you could buy a lower-end socket AM3+ processor like the FX4350, otherwise Xeon is clearly the better choice.

What most people don't realize is that the desktop version is basically an afterthought. The chip has been optimized for laptops, where it does make some sense (adding a discrete GPU is not an option after purchase and laptops with discrete GPUs are quite more expensive, so the comparative advantage is more important). AMD knows they can't win on the desktop, which is why they didn't bother with extreme caches, 4-module (8-core) versions and cherry-picked chips with crazy TDPs. Personally, I'm much more excited with the laptop version of Kaveri, such as the 7350B in the HP EliteBook 745 G2.

Anyway, for the price it makes a really great casual gaming PC, especially for people who are price sensitive and can't afford a +$100 discrete GPU (in some places this is a decent chunk of a month's salary...).

I don't really see the need for gaming on the go, and if such a need exists, isn't it sufficiently covered by existing gadgets (smartphone or non-gaming tablet)? Furthermore, are the current tablet games worth buying hardware specifically for the job? What would be the point of Angry Birds at 120fps?

Anyway, the hardware looks cool, but the fact that no other manufacturer bothered to use the nvidia hardware is a bit disconcerting. If it were the best thing since sliced bread, many designs would have flooded the market.