Is there some kind of dread Dale Carnegie template that gets used to construct all of these presentations? Is it the one entitled: "Bore your prospects to tears"? Because it's the only explanation I can think of.

Basically, all these presentations are done backwards. Point three in the SK Murphy list is the most important one for me:

"'Why I should care about your product' is left to the end of the presentation."

There are two problems with this. Until I got bored with doing it, I could happily chip away at the setup slides to the extent that some presentations simply ran out of time. And it doesn't help get a story written because all the information is coming in reverse order.

The problem with the time-share pitch is that it is designed to sell. Unless you're selling pens, it's pretty unlikely that the hack is going to buy one. (OK, it's different for gadget makers but the launches I've seen for those have largely been "look, shiny stuff").

At regular intervals in a briefing, I will be thinking: "What's my best possible introduction for a story?" And, most of the time, the answer will be: "I wish I knew."

People really need to think about the thought processes that their intended audience are likely to use. A journalist is, in the case of a briefing, looking for a story. They may well not take away the story you presented but if you start off with what you think the story is, things might at least unfold in the right order.

So, make the claim early. And then provide the background for why this claim might be true. And then you can move onto the background. Why this way round? Because it's a structure that fits the inverted pyramid of news; it fits the thought processes that journalists are most likely to use: what's happened; how it happened; evidence to back it all up.

Readers of news stories do the same thing: they want to know what happened. And, if they are still interested, they stick around to find out about the hows and the whys.

Ideally, I'd ditch the slides and go straight for an interview in which the questions follow this rough sequence and expressed in their most brutal form: what have you done that you think is so great; how is this different to what has gone before; how did you set about doing this; why should people think you have experience here/what's your background?

I suspect it's a structure that will work just as well for VCs and sales prospects in the technology business.

I've vowed to try to keep it useful, doing things such as flagging up features I'm working on. We'll see how long that lasts.

"Larrabee's highly threaded x86 architecture benefits traditional enterprise throughput computing applications, such as text indexing [Intel's emphasis]."

I'm not sure when text indexing became a "traditional enterprise application" but it's precisely what search-engine operators like to do a lot of. Maybe it's no coincidence that one of the researchers who wrote on asymmetric multiprocessing in IEEE Computer this month is now at Google.

As far as 3D graphics goes, it looks as though Intel has taken a similar approach to ARM to try to reduce the amount of memory bandwidth issue in 3D graphics. It is using tile-based algorithms to split up the processing rather than the immediate mode favoured by high-end GPUs. The caches can also be loaded by software - their cache behaviour can be overridden, which should help in 3D work.

Long term, an underperforming Larrabee may not be enough to preserve the graphics processor market for nVidia. Larrabee's architecture is a hint of what a future x86 processor will look like. If Intel gets its way, much of the processing that goes on today in a dedicated GPU will be sucked into the core processor. That Intel decided to go with plonking down loads of reworked Pentium cores onto the Larrabee die rather than a new processor indicates that the company has its eye on the future of the x86 rather than today's GPU market.

Intel has concentrated on programming ease by giving each processor regular level-one and level-two caches instead of software-managed local memories. Caches are good for programmers but they can get in the way when the emphasis is on making the most of memory bandwidth with graphics-intensive work. And memory bandwidth is precious on a GPU, particularly as power consumption becomes more of an issue.

"The Larrabee architecture fully supports IEEE standards for single and double precision floating-point arithmetic. Support for these standards is a pre-requisite for many types of tasks including financial applications."

The approach taken by Intel fits a similar pattern to the i860, a processor from the early 1990s that started life as an experiment and became, briefly, one of the most popular processors in parallel supercomputers and high-end 3D graphics rendering. It turned up in the NeXT workstation and the Silicon Graphics Onyx Reality Engine.

Intel has never been that successful with discrete graphics. And it has had a few goes. The purchase of Chips & Technologies in the late 1990s did not go anywhere - it was at that time that ATI and nVidia came to the fore, seeing off Intel and companies such as 3DFX. The only way that Intel has made inroads is by integrating graphics into its chipsets. Even then, a lot of standalone GPU sales are into PCs with integrated graphics.

Over time, however, software wins. The GPUs will gradually acquire more of the features of host processors. And Intel will have enough spare transistors come the 22nm generation to pull something like a Larrabee onto the main die, coupling it with maybe four or eight big processor cores in an asymmetric multiprocessor. In the next few years, Intel gets to work out how the two bits would talk to each other, so you could regard the design of Larrabee as an architectural shift.

You could then strip the GPU down to its essential parts: providing hardware for the things that are just too expensive to do in software. What hardware goes onto the standalone GPU will depend on whether Intel is right in thinking ray-tracing will take over or, more likely, games programmers will use hybrids of today's techniques and more processor-intensive algorithms. Other programmers can use interfaces such as OpenCL to make use of the extra cores if they don't want to program them directly.

In this environment, Intel does not need to be successful in graphics per se. It simply has to be able to convince graphics programmers that their long-term future lies in much more software-intensive rendering. AMD could follow a similar path having bought ATI. If it does, and splits the GPU in two, then Intel's job gets easier. And nVidia's task is much tougher.

What's the problem? Just take the processor core you have already and then step and repeat it across the die. It's worked for graphics processors.

Unfortunately, only some software parallelises so well that it will spread across many cores. Many times, the overhead of distributing the work outweighs the advantage you get from running the code in parallel. This, in effect, is the modification that Gene Amdahl made to his eponymous law of performance in computers.

In its most basic form, Amdahl's Law says it's only worth speeding up things that you do a lot. Big, nested loops are good targets. Lots of branching straight-line code? Not worth the effort. With parallel processors, if you can spread the work of loops over many of them, you see a speed-up. But there is a limit governed by how much code you need to run on just one processor.

In a paper published in this month's IEEE Computer, a pair of researchers from the University of Winsconsin-Madison - one of whom has now moved to Google - has attempted to extend Amdahl's Law to the world of multicore processors where you do not necessarily make all the processors the same size.

Borkar cited Pollack's Rule, which says the performance increase of a processor is roughly proportional to the square root of the increase in its complexity. Put another way, if you double the number of transistors in a processor, the chances are that you will only get a speedup of about 40 per cent.

In this environment, having lots of fairly simple cores looks to be the favourite option. But, power becomes a big, big problem. Borkar noted that simply sticking slavishly to this plan would have a devastating impact on power consumption, because all of the energy is used to support the interconnect that feeds all the processors.

Using 100 much more complex, bigger cores would bring the power needed for the mesh down by a factor of ten to just 15W. "What we need is a really careful balance," he said. "Just because we can integrate thousands of cores, don’t get carried away and implement a thousand cores."

Tilera's Anant Agarwal defined his own version of Pollack's Rule: the Kill Rule. This says: "A resource in a core must be increased in area only if the core’s performance improvement is at least proportional to the core’s area increase."

Agarwal reckoned that dual-issue superscalar processors will become the norm as they largely satisfy his Kill rule, although there is a broad spread of those designs. A complex dual-issue architecture might need 15 million transistors, including cache. The simpler designs offered by companies such as ARM and MIPS Technologies, which rely more on the compiler’s ability to schedule instructions effectively, might be squeezed into a 5 million transistor block.

The argument from Mark Hill and Michael Marty from the work at Winsconsin-Madison is that you can do it all. With enough transistors to play with, you can have some brainiac superpipelined, superscalar monster that gets to work on all the straight-line code and have lots of smaller, simpler processors get on with the parallelisable stuff.

They show, for a very large array, the monster-plus-servants approach shows reasonable acceleration for code that is 90 per cent or more parallelisable and certainly performs better than the situation where all cores are equal.

There is a warning here. This assumes that your only constraint is the number of transistors. It doesn't take account of interconnect overhead or power consumption. But it's a result that tends to vindicate the asymmetric design of the modern PC if you start to use the GPU for acceleration.

Where the gains could be greater is in the idea of a reconfigurable array. The idea is that you make a monster processor out of lots of little ones. The speedups on this look great. Then again, given that nobody knows how to make this kind of machine, you would hope so.

The problem is that big processors are not made out of little ones. It's more that, inside every monster processor there is a little one and a chunk of stuff that tries valiantly to keep shoving data through that processor. You can't easily turn processors into prediction and speculation units, at least not with current architectures.

Hill and Marty argue that, despite this, it is a result that is worth closer inspection. What if you could reorganise processors to act as speculation units? Or maybe come up with new architectures that can be split apart and recombined as needed? There is one architecture that looks as though it could be a contender for a separable architecture: this is the TRIPS processor out of the University of Texas. It starts out as an array of processors with an instruction-scheduling scheme that pulls them into virtual machine.

TRIPS calls for a change to software compilation. You also have the work on speculative compilers such as Codeplay's Sieve-C, which runs on regular combinations of PC processors and GPUs. This is designed to work on parallelisable code - it expands the amount of software that can be run in parallel by allowing it to make mistakes and then fixing them afterwards. This kind of strategy might be used to have all the speculation work performed in today's monster processors done in software on a bunch of simple processors.

There are no ready answers, but Hill and Marty have indicated some areas where it might be worth computer architects looking, just as long as the performance predictions aren't completely knackered the minute you invoke power consumption or interconnect overhead.

It's nothing more than saying the company has a white paper on how funky femtocells will be. But, there is a chance there could be something useful in it and I was just beginning to line up interviews. So, I first went to the link to download the file...and found out I need to fill in a form to get it. I'm not over-fussed about filling in a form but, as this is likely to go into some form of CRM system, I figure it's just as easy to save a salesperson a call and get it from the PR. Who, of course, will be named at the bottom of the release and set up the interview at the same time.

I hit Reply and start banging out the email only to notice an odd bit of text in the introduction "...please do not hesitate to contact them via the details below". Then I realise that Neondrum, who sent the release is only distributing the release, with all the usual disclaimers: "[We] cannot accept any liability whatsoever for the inaccuracy or otherwise of any information contained in this news release" etc. All they're going to do is tell me to contact the client. OK fine.

But there are no other contact details.

I asked Neondrum to send them over. Ten hours later (in the meantime, I'd found that CompanyCare had been looking after Ip.access and contacted them directly) I got a reply:

"Sorry, the introductory message was badly worded - ip.access haven't provided a specific contact for this media advisory, if you want to find out more you need to download the paper."

Thanks. That's so helpful. And from a company that publishes a booklet that it claims contains ten top tips for online PR. I wonder if "always provide a contact number" is in there. I'd find out, but you have to register as a client at Neondrum to stand a chance of getting it. If someone has a copy, send it over, I could do with a laugh.

A few hours later, the Cuil site died. Oops. But, no mind, just the effect of thousands of people hitting the site to see how it performed versus Google. "Flatlining right after your launch is more of a rite of passage than an embarrassment."

A day later and the euphoria had gone. "The story quickly turned from Google killer to Google's lunch."

Getting a backlash so quickly? "This was entirely the company's own fault. It pre-briefed every blogger and tech journalist on the planet, but didn't allow anyone to actually the test the search engine before the launch," complained Erick Schonfeld.

And you're surprised? Who says the old promotional tricks don't work?

Cuil got the backlash in early. In the old days, this kind of backlash would take weeks. It took just one fishwrap cycle for Cuil. But, this arguably gives the company the chance to recover or to simply fade away. Whatever they do next, there are a lot of people who will remember who they are. Can you say the same for Hakia or any of the other Google wannabes? (For what it's worth, I don't think any the putative Google replacements are going to get anywhere - the technology cycle doesn't work like that. Like rarely gets replaced with like unless the leader stumbles really badly. It's generally a different concept that shoves the old guard out of the way.)

A similar thing happened with the 'Ginger/Segway' launch, although this one played out in a different way. There was a lengthy "Ginger is coming. It will change the world" phase, helped along because the company concentrated on Dean Kamen versus the product. Then the Segway launched and everybody laughed. I still think it's a joke machine but they keep making them. And you don't find that many people who haven't heard of a Segway.

Tom at Tom's Tech Blog points to Sarah Lacy's criticism of the rush to publish that made so many writers publicly change their minds in less than a day. Her argument is that people should hang back, think a bit, then publish. That's a lovely thought. It is one I shall treasure. Because, although she's right on that score, in principle. Unfortunately, it doesn't work.

The current ad-driven business model for Internet publishing does not reward delay, no matter how good the analysis is. The initial burst of interest is the one publishers want to catch: waiting does not often get you a big audience. And, although the marketing consultants bang on about 'conversation' and 'engagement', the sites making money off of publishing need audience to sell ads. PRs also know this and exploit it: that is what they are paid to do. It's very easy to forget while doing interviews, especially under embargo - I know that I forget it all the time - but you always have to be thinking: "What have you kept hidden and why?"

For every piece of placebo-tech in the briefing - "it's got relativity powered indexing for improved relevancy scoring" - there's a caveat. Spotting the caveats is, however, more than half the fun.

Nokia decided to outsource much of its silicon design and the company has traditionally bought the off-the-shelf ARM cores anyway, which Texas Instruments then put into system-on-chip (SoC) parts. Motorola already has an architectural licence. Samsung would be a possibility as the world's second largest chipmaker, but signed a big deal with ARM earlier this year to get early access to ARM's own designs.

Following Apple's purchase of chip design firm PA Semiconductor, ARM people have been particularly jumpy of late whenever Apple gets mentioned. And questions asking whether Apple already has an architectural licence (the computer maker was the driving force behind the creation of ARM and one of the original investors) were met with a "you'll have to ask Apple", rather than a "yes", a "no" or a "no comment".

PA's designers have a lot of experience with ARM, although their most recent offering, which is getting dropped like a stone, was based on the PowerPC. ARM's investor meeting is about to start. But, realistically, if the company was going to say that Apple is the new architectural licensee, it would have done so already.

ARM CEO Warrren East just warned that it will take time for the company to see royalties from any products sold that use the processors licensed under the new arrangement (9:20): "It is an architectural licence with a leading OEM for both current and future technology. Don't get too excited on any revenue on this: it will take some time. The revenue [from this deal] will be recognised over several years."

ARM's emphasis on this deal is that it is all about the future. Tim Score, CFO, said (9:38): "When ARM signs architectural licences, they are typically for an architecture that is already in play. So you tend to get a big revenue bump. This one is also for future architectures, so the revenue has to be spread over a number of years."

Knol is a magnet for the get-rich-quick brigade who reckon they can siphon off a load of money through ads for dodgy health supplements. It might work as a competitor to Squidoo, Mahalo and even Wikipedia. But, a lot will depend on the image that Knol attracts in the short term. It's got a good chance of becoming the .info of information and quick reference sites, where the only people who show up are spammers with slightly more original content.

But does that matter to Google? Regarding Google's business as being in search is a mistake. It's an advertising business. And one of the unfortunate drivers of the online classified ad business that the company now effectively dominates is that a bunch of people are only too happy to click on ads for the 'health supplements', teeth whiteners and other kind of products being actively promoted on Knol pages. They may well be the most active ad-clickers around.

There's a good chance that Google will make more cash out of the dodgier Knol pages than the ones designed to look more like entries in an encyclopedia.

People are misreading Google's slogan, "Don't be evil". It's not a slogan. It's an admonishment to those sucking on the Adsense teat: "Don't be evil...or we'll kick you off the search results pages. You can be a bit naughty, mind."

While things are good for Google, nobody will really care:

"You show them you have in you something that is really profitable, and then there will be no limits to the recognition of your ability. Of course you must take care of the motives - right motives - always." - Mr Kurtz, Heart of Darkness

Why isn't this their marketing slogan?

I don't know. Because the policy won't last?