Slashdot Mirror

with the current AI revolution and quantum computing can also match the mental capabilities of most humans as well.

The current AI 'revolution' isn't going to give us computers with mental capabilities that match most humans. If you're really interested in understanding the state of AI, Peter Norvig is doing a town hall presentation, I encourage you to visit it.

I would view this ACM webinar on the five sub-disciplines in machine learning (assuming you can) and then investigate the listed resources:

        http://event.on24.com/wcc/r/10...

Quick note: This article is a spin off of what Eric had to say during the most recent Black Hat Webcast, where Jeremiah Grossman was talking about clickjacking and other related browser issues. Eric made a lot of sense talking about plug ins and addons being the cross platform low hanging fruit.

Listen and watch the webinar to hear what he had to say and keep everything in context:
http://w.on24.com/r.htm?e=122494&s=1&k=05ED21C1734D531D2D84CA56F4ADB0F2

Or download the .m4b audio file when we get it online next week here:
https://www.blackhat.com/html/webinars/webinars-index.html

Not until superconductors are workable. Even in the short runs between servers, DC tends to have higher losses. You're better off spending money on higher efficiency power supplies.

http://event.on24.com/event/95/75/4/rt/1/documents/player_docanchr_1/wp63_fr.pdf

interesting.
I still am skeptical about static checking be very effective in seeing race conditions...but then Coverity also as dynamic checking tools: https://event.on24.com/eventRegistration/EventLobbyServlet?target=registration.jsp&eventid=110122&sessionid=1&key=678B7E9A4BA76B1FB9659E52F0946162&partnerref=5SD&sourcepage=register

Their benchmarks are already flawed. If you watched the launch, they said it SEVERAL times, RAM is running at 1ghz.

Here's the link, it's a 1.5 hr video with slides, but was worth EVERY minute. Some NICE examples of how powerful the card is.

Click Here to see presentation

Good results

First, we have the second quarter fiscal 2002 results, released last week. These were actually really quite good. VA has reduced its cash burn to $6.1m/quarter this is not only a massive fall from the hardware services days of a >$30m cash burn, but is substantially below the target of $8m/quarter which VA announced at the time of quitting the hardware business. Having left the hardware and consulting businesses, VA was concentrating on selling its main software product, Sourceforge 3.0, and had made a number of new sales to blue-chip customers such as Stanford Universty and Pfizer. We had a few problems with their statement in the conference call and the press release that they had "$61m in cash and marketable securities" which is true, but highly misleading as to their actual financial position as they also have current liabilities of $18m (ie; they need this much to pay bills falling due in the next six months, so the actual cash available to burn is more like $43m), and we regard their description of the redundancy payments and lease cancellation fees which make up their restructuring costs as "non cash items" as actively ludicrous, but this is nit-picking; the facts as of a couple of weeks ago appeared to be that VA Software was on the raspberry road to profitability.

But .

Then we got this little bombshell; Slashdot, jewel in the crown of VA Software's OSDN network of Open Source websites, is moving to a pay subscriptions model a la Salon. Well, perhaps that's being a little bit too harsh; Slashdot isn't doing the full reader reduction exercise of making you pay for the only content you came to read, but it is going to be having "more intrusive" ads (by which I think we mean expanded banners and skyscrapers surely Slashdot wouldn't dare to go down the route of pop-ups or interstitials, would they? WOULD THEY? AARRGH!), and you'll be able to view slashdot without these ads at the bargain subscription rate of $5 per thousand pages. Obviously, this caused much wailing and gnashing of teeth among the assorted slashbots (2275 comments so far, nearly a third as many as adequacy.org's most popular article), but we can't help thinking that they're missing the point. Nobody, least of all VA, thinks that there will be material revenue opportunities from the subscription model; all this is, is a figleaf designed to allow Slashdot to accept pop-up X10 ads while giving its editors hobbyists Rob Malda and Jeff Bates, a lightning conductor of "well, why don't you subscribe?" to deal with the floods of threatening email they are likely to receive.

So fair enough. But when we read the actual announcement on Slashdot, we at adequacy.org got worried. When we think we're looking at a company which is on the right track, we don't like to see senior staff at its only profitable business unit making statements like:

" The large ads that you see on many other sites are coming here. We really don't have an option: these are what advertisers want, and if we don't provide them, we won't be around much longer"

or

" We won't create subscriber only features that cost more to maintain than they generate. But we do need support from you if we are to continue."

What the hell? Slashdot was known to be profitable and cash positive when taken over by Andover.net in 1999. Andover.net was known to be profitable and cash-positive when taken over by VA Linux in 2000. The OSDN group of sites was, according to the 2Q02 results conference call, the source of more or less all the revenue generated by VA Software. And now we're being told that the ad market is so precarious that the VA cash pile is likely to be burnt up imminently? What gives? Quite apart from anything, statements like "we won't be around much longer" are Forward Looking Statements. Companies with publicly traded securities outstanding should not be making forward looking statements outside of the context of a scheduled conference call or an announcement to the general public under Regulation FD. It is, quite simply, not good enough for Rob Malda to be making this kind of wild assertion about the trading conditions faced by the key media property in the only profitable division of VA Software, ad hoc and without any kind of "safe harbor" statement. We don't know whether or not this announcement was technically in breach of Section 21 (E) of the Securities Exchange Act 1934, but we do know that well-managed companies with competently run press office and investor relations functions don't leak rumors in this kind of way.

Adequacy investigates

When we at adequacy.org witness an informational cluster-fuck like this in the making, we want to dig and delve, for the benefit of you our readers. We're about to make a few fairly controversial statements in this report, and we'd like you to take the following on trust: all the statements we make below which are in bold face can be sourced to a prominent (as in, you'd recognise the name if we told you) employee of Slashdot. We at adequacy don't want to cost anyone their job, so we'll make the following statement:

The statements sourced to an employee of Slashdot were acquired as the result of IRC conversation on an open channel. For this reason, adequacy.org does not feel bound to protect its source come what may. However, on general principles, we will only hand over the IRC logs which prove the veracity of our information on receipt of a subpoena from VA Software. In the event of our receiving such a subpoena, we will do our very best to publicise throughout the Internet the fact that VA Software issued such a subpoena to us in order to track down a critical employee, something which we would imagine would not generate good publicity with the core slashdot audience.

Sourceforge is not profitable and looks like it never will be. According to our source, "it's a giant vacuum". And this seems about right to us. The recent conference call with VA Chief Executive Larry "Eleven Million Dollar Man" (that's how much VA stock he's sold for cash since the float) Augustin was full of the joys of Sourceforge "Enterprise Edition" 3.0, a "proprietary" version of the popular Open Source collaborative software development tool. Indeed, in response to a question, VA's Chairman and Chief Executive told the world that VA Software (a company which, according to its CFO made "substantially all" of its revenue from the online advertising of the OSDN) was "a company in the enterprise software market". Much was made of the fact that new sales had been made to Stanford and Pfizer, two new key clients. But when you try to pin down these sales to hard revenue numbers, it kind of drifts away. The hard fact is that Sourceforge charges $1000 per seat license (there are apparently issues relating to revenue recognition over the term of the long-term licensing contracts which VA is trying to sll, but $1K was the hard number given at the conference call). That means that, before VA Software can be considered to be mainly a software company, it needs to be selling 5000 seats worth of Sourceforge per quarter (generating $5m of revenue, roughly the same as OSDN's revenue). How close is it now to that goal?

Not close. Although the reference implementation of Sourceforge; the licensing level at which it starts to generate positive RoI for its customers, is estimated to be 120 seats, the vast majority of its current customer base are installing it on trial implementations of 30 seats to see if it's any good. Two or three big sales of Sourceforge might make a quarter of a million bucks at the outside; Sourceforge revenue for 2Q02 might possibly be as low as $60,000. Since Sourceforge 3.1, with better integration with other tools and added functionality is on the way, we can't see anyone springing for a full installation of 3.0, meaning that sales are at the mercy of the development schedule. In any case, we're not sure why anyone would buy 3.0; as far as we can tell, the main advantage over the Open Source version is that you get to use Oracle rather than PostGreSQL as a back end, which shouldn't be too terribly hard an alteration to make in-house given that the source code for the biggest existing implementation of Sourceforge (http://www.sourceforge.net) is available.

So, on the basis of publicly verifiable facts, our source appears to know what he's talking about.

OSDN is run tightly; VA as a whole is not. This is more or less a direct quote from our source, and we believe it. OSDN, for all its expensive branding and new name, is the business of Andover.net, which was always the poor man's CMG, or Ziff-Davis for the technologically literate. Which is to say, a bunch of guys who knew how to sell ads for computer stuff. They're still good. Let's consider the following:

Again from the conference call, we learn that in 2Q02, Intel accounted for 20% of total revenues. That's (cue drum roll, Dr Evil voice) one million dollars! Did they buy a thousand Sourceforge seats? To put it bluntly, no. They spent this on advertising

You can't spend one million dollars on advertising

At any reasonable CPM rate (or indeed, at OSDN's quoted rates for "selfserve" ads recently posted, one million dollars would buy you 250 million ad impressions. According to the OSDN advertising screen, they serve 120 million page views a month. So, by this standard, roughly two out of every three ads on OSDN during the second quarter of fiscal 2002 would have been ads for Intel. I have to tell you, and every regular viewer of Slashdot will agree, that they weren't.

Slashdot is notorious for running ads for thinkgeek tshirts, other OSDN sites and caffeinated mints, but surprisingly few ads for the high-end server gear which is the unique selling point of OSDN to its advertiser base. And slashdot accounts for an awful lot of those 120 million pages. Specifically, according to figures given in in Malda's statement, Slashdot has "one third of a million visitors per day", and the median visitor generates ten pageviews (we guesstimate this from the statement that, at a subscription rate of $5 per 1000 pages without ads, "82% of our readers could view slashdot for a year for $20", ie, 4000 pages per year). That means that over a quarter, just about 90 million of OSDN's 120 million pages are accounted for by Slashdot. So if Intel has spent One Million Dollars on OSDN advertising without making a material impact on slashdot, then something pretty strange has gone on.

Here's our guess. Intel is the sponsor of the "Large Linux Installation Foundry" on sourceforge.net. What's been going on here is "narrowcasting" Intel isn't so much interested in serving 250 million pages to random Slashbots, but is more interested in serving about 400 pages over the quarter to a group of people possibly as small as nine or ten, who were making the decision in 2Q02 about which technology provider they would be going for in a large Linux installation. It is not at all unknown for big ticket computer salesmen to drop a seven-figure check in promotions if they're hoping to land a nine-figure contract. It's also not impossible that the sponsorship of Sourceforge Large Linux Installations during 2Q02 was the subject of a bidding war between to rivals over the same large contract. We can't prove this, but we're pretty sure that something of this sort happened (if there are any more disgruntled VA employees out there, we'd love to know if we were right). In any case, it's not what you might call "high-quality income"; although VA hope to continue doing business with Intel, this is a big chunk of revenue to be dependent on one piece of marketing whim.

Slashdot could be sold to another media organisation. We had to read between the lines to get to this one, and it's probably not fair to pin it on our source, but he certainly entertained our speculation on the subject. And the interesting thing is that, with the information we were able to glean about the decomposition of 2Q earnings, Slashdot doesn't look like the cash cow for VA that we thought it might be. Out of the $5m revenue of VA Software, we can take out approximately $750K of interest income on the cash balance and maybe $200K for Sourceforge, meaning that the Intel contract accounts for roughly a quarter of the operating income of OSDN. From the pagecount, we know that Slashdot accounts for three quarters of the pageviews (and thus roughly three quarters of the bandwidth costs); to assume that it generates three quarters of the revenue would be tantamount to assuming that the other OSDN sites make next to zero revenue. Which is a crazy assumption, particularly given the intangible benefit to VA Software of having sourceforge.net as a promotional device for Sourceforge Enterprise Edition. And if Slashdot accounts for three quarters of the costs and less than three quarters of the revenues, it's a dog in the OSDN portfolio, not a star or a cash cow.

So, why not sell it? Although Slashdot may be a drain on the average profitability of OSDN, it probably breaks even, and in the world of magazine publishing, that's not bad. Publishing companies know that profitability has to be measured across a portfolio of magazines, not unit by unit, and it's often worth your while publishing a loss-making Talk Magazine for a while for the touch of stardust glamour it adds to a lucrative (but potentially rather prosaic) Conde Nast Traveller. Slashdot would be a perfect "hood ornament" for a profitable stable of computer magazines, dragging the kids in while they were in college and then cross-promoting them onto other titles by the time they had reached a saleable demographic. And all this could be done without compromising its "editorial integrity", which is something usually respected in the media world, though not so much in the software publishing world ("Andover.net had all sorts of evil plans for Slashdot", our source reveals).

Bottom line: If Larry Augustin wants to claim to be running a company in the enterprise software business, it's time for him to walk the talk. Let's see some divestment of non-core assets like Slashdot. Otherwise, we ought to be facing facts and reminding ourselves that the company which used to be "VA Linux" and is now "VA Software" has always been "VA Media". It's a publishing company, and ought to be managed as one. If that means getting rid of Eleven Million Dollar Larry and getting a graduate of the Si Mewhouse academy, then so be it.

Good results

First, we have the second quarter fiscal 2002 results, released last week. These were actually really quite good. VA has reduced its cash burn to $6.1m/quarter this is not only a massive fall from the hardware services days of a >$30m cash burn, but is substantially below the target of $8m/quarter which VA announced at the time of quitting the hardware business. Having left the hardware and consulting businesses, VA was concentrating on selling its main software product, Sourceforge 3.0, and had made a number of new sales to blue-chip customers such as Stanford Universty and Pfizer. We had a few problems with their statement in the conference call and the press release that they had "$61m in cash and marketable securities" which is true, but highly misleading as to their actual financial position as they also have current liabilities of $18m (ie; they need this much to pay bills falling due in the next six months, so the actual cash available to burn is more like $43m), and we regard their description of the redundancy payments and lease cancellation fees which make up their restructuring costs as "non cash items" as actively ludicrous, but this is nit-picking; the facts as of a couple of weeks ago appeared to be that VA Software was on the raspberry road to profitability.

But .

Then we got this little bombshell; Slashdot, jewel in the crown of VA Software's OSDN network of Open Source websites, is moving to a pay subscriptions model a la Salon. Well, perhaps that's being a little bit too harsh; Slashdot isn't doing the full reader reduction exercise of making you pay for the only content you came to read, but it is going to be having "more intrusive" ads (by which I think we mean expanded banners and skyscrapers surely Slashdot wouldn't dare to go down the route of pop-ups or interstitials, would they? WOULD THEY? AARRGH!), and you'll be able to view slashdot without these ads at the bargain subscription rate of $5 per thousand pages. Obviously, this caused much wailing and gnashing of teeth among the assorted slashbots (2275 comments so far, nearly a third as many as adequacy.org's most popular article), but we can't help thinking that they're missing the point. Nobody, least of all VA, thinks that there will be material revenue opportunities from the subscription model; all this is, is a figleaf designed to allow Slashdot to accept pop-up X10 ads while giving its editors hobbyists Rob Malda and Jeff Bates, a lightning conductor of "well, why don't you subscribe?" to deal with the floods of threatening email they are likely to receive.

So fair enough. But when we read the actual announcement on Slashdot, we at adequacy.org got worried. When we think we're looking at a company which is on the right track, we don't like to see senior staff at its only profitable business unit making statements like:

" The large ads that you see on many other sites are coming here. We really don't have an option: these are what advertisers want, and if we don't provide them, we won't be around much longer"

or

" We won't create subscriber only features that cost more to maintain than they generate. But we do need support from you if we are to continue."

What the hell? Slashdot was known to be profitable and cash positive when taken over by Andover.net in 1999. Andover.net was known to be profitable and cash-positive when taken over by VA Linux in 2000. The OSDN group of sites was, according to the 2Q02 results conference call, the source of more or less all the revenue generated by VA Software. And now we're being told that the ad market is so precarious that the VA cash pile is likely to be burnt up imminently? What gives? Quite apart from anything, statements like "we won't be around much longer" are Forward Looking Statements. Companies with publicly traded securities outstanding should not be making forward looking statements outside of the context of a scheduled conference call or an announcement to the general public under Regulation FD. It is, quite simply, not good enough for Rob Malda to be making this kind of wild assertion about the trading conditions faced by the key media property in the only profitable division of VA Software, ad hoc and without any kind of "safe harbor" statement. We don't know whether or not this announcement was technically in breach of Section 21 (E) of the Securities Exchange Act 1934, but we do know that well-managed companies with competently run press office and investor relations functions don't leak rumors in this kind of way.

Adequacy investigates

When we at adequacy.org witness an informational cluster-fuck like this in the making, we want to dig and delve, for the benefit of you our readers. We're about to make a few fairly controversial statements in this report, and we'd like you to take the following on trust: all the statements we make below which are in bold face can be sourced to a prominent (as in, you'd recognise the name if we told you) employee of Slashdot. We at adequacy don't want to cost anyone their job, so we'll make the following statement:

The statements sourced to an employee of Slashdot were acquired as the result of IRC conversation on an open channel. For this reason, adequacy.org does not feel bound to protect its source come what may. However, on general principles, we will only hand over the IRC logs which prove the veracity of our information on receipt of a subpoena from VA Software. In the event of our receiving such a subpoena, we will do our very best to publicise throughout the Internet the fact that VA Software issued such a subpoena to us in order to track down a critical employee, something which we would imagine would not generate good publicity with the core slashdot audience.

Sourceforge is not profitable and looks like it never will be. According to our source, "it's a giant vacuum". And this seems about right to us. The recent conference call with VA Chief Executive Larry "Eleven Million Dollar Man" (that's how much VA stock he's sold for cash since the float) Augustin was full of the joys of Sourceforge "Enterprise Edition" 3.0, a "proprietary" version of the popular Open Source collaborative software development tool. Indeed, in response to a question, VA's Chairman and Chief Executive told the world that VA Software (a company which, according to its CFO made "substantially all" of its revenue from the online advertising of the OSDN) was "a company in the enterprise software market". Much was made of the fact that new sales had been made to Stanford and Pfizer, two new key clients. But when you try to pin down these sales to hard revenue numbers, it kind of drifts away. The hard fact is that Sourceforge charges $1000 per seat license (there are apparently issues relating to revenue recognition over the term of the long-term licensing contracts which VA is trying to sll, but $1K was the hard number given at the conference call). That means that, before VA Software can be considered to be mainly a software company, it needs to be selling 5000 seats worth of Sourceforge per quarter (generating $5m of revenue, roughly the same as OSDN's revenue). How close is it now to that goal?

Not close. Although the reference implementation of Sourceforge; the licensing level at which it starts to generate positive RoI for its customers, is estimated to be 120 seats, the vast majority of its current customer base are installing it on trial implementations of 30 seats to see if it's any good. Two or three big sales of Sourceforge might make a quarter of a million bucks at the outside; Sourceforge revenue for 2Q02 might possibly be as low as $60,000. Since Sourceforge 3.1, with better integration with other tools and added functionality is on the way, we can't see anyone springing for a full installation of 3.0, meaning that sales are at the mercy of the development schedule. In any case, we're not sure why anyone would buy 3.0; as far as we can tell, the main advantage over the Open Source version is that you get to use Oracle rather than PostGreSQL as a back end, which shouldn't be too terribly hard an alteration to make in-house given that the source code for the biggest existing implementation of Sourceforge (http://www.sourceforge.net) is available.

So, on the basis of publicly verifiable facts, our source appears to know what he's talking about.

OSDN is run tightly; VA as a whole is not. This is more or less a direct quote from our source, and we believe it. OSDN, for all its expensive branding and new name, is the business of Andover.net, which was always the poor man's CMG, or Ziff-Davis for the technologically literate. Which is to say, a bunch of guys who knew how to sell ads for computer stuff. They're still good. Let's consider the following:

Again from the conference call, we learn that in 2Q02, Intel accounted for 20% of total revenues. That's (cue drum roll, Dr Evil voice) one million dollars! Did they buy a thousand Sourceforge seats? To put it bluntly, no. They spent this on advertising

You can't spend one million dollars on advertising

At any reasonable CPM rate (or indeed, at OSDN's quoted rates for "selfserve" ads recently posted, one million dollars would buy you 250 million ad impressions. According to the OSDN advertising screen, they serve 120 million page views a month. So, by this standard, roughly two out of every three ads on OSDN during the second quarter of fiscal 2002 would have been ads for Intel. I have to tell you, and every regular viewer of Slashdot will agree, that they weren't.

Slashdot is notorious for running ads for thinkgeek tshirts, other OSDN sites and caffeinated mints, but surprisingly few ads for the high-end server gear which is the unique selling point of OSDN to its advertiser base. And slashdot accounts for an awful lot of those 120 million pages. Specifically, according to figures given in in Malda's statement, Slashdot has "one third of a million visitors per day", and the median visitor generates ten pageviews (we guesstimate this from the statement that, at a subscription rate of $5 per 1000 pages without ads, "82% of our readers could view slashdot for a year for $20", ie, 4000 pages per year). That means that over a quarter, just about 90 million of OSDN's 120 million pages are accounted for by Slashdot. So if Intel has spent One Million Dollars on OSDN advertising without making a material impact on slashdot, then something pretty strange has gone on.

Here's our guess. Intel is the sponsor of the "Large Linux Installation Foundry" on sourceforge.net. What's been going on here is "narrowcasting" Intel isn't so much interested in serving 250 million pages to random Slashbots, but is more interested in serving about 400 pages over the quarter to a group of people possibly as small as nine or ten, who were making the decision in 2Q02 about which technology provider they would be going for in a large Linux installation. It is not at all unknown for big ticket computer salesmen to drop a seven-figure check in promotions if they're hoping to land a nine-figure contract. It's also not impossible that the sponsorship of Sourceforge Large Linux Installations during 2Q02 was the subject of a bidding war between to rivals over the same large contract. We can't prove this, but we're pretty sure that something of this sort happened (if there are any more disgruntled VA employees out there, we'd love to know if we were right). In any case, it's not what you might call "high-quality income"; although VA hope to continue doing business with Intel, this is a big chunk of revenue to be dependent on one piece of marketing whim.

Slashdot could be sold to another media organisation. We had to read between the lines to get to this one, and it's probably not fair to pin it on our source, but he certainly entertained our speculation on the subject. And the interesting thing is that, with the information we were able to glean about the decomposition of 2Q earnings, Slashdot doesn't look like the cash cow for VA that we thought it might be. Out of the $5m revenue of VA Software, we can take out approximately $750K of interest income on the cash balance and maybe $200K for Sourceforge, meaning that the Intel contract accounts for roughly a quarter of the operating income of OSDN. From the pagecount, we know that Slashdot accounts for three quarters of the pageviews (and thus roughly three quarters of the bandwidth costs); to assume that it generates three quarters of the revenue would be tantamount to assuming that the other OSDN sites make next to zero revenue. Which is a crazy assumption, particularly given the intangible benefit to VA Software of having sourceforge.net as a promotional device for Sourceforge Enterprise Edition. And if Slashdot accounts for three quarters of the costs and less than three quarters of the revenues, it's a dog in the OSDN portfolio, not a star or a cash cow.

So, why not sell it? Although Slashdot may be a drain on the average profitability of OSDN, it probably breaks even, and in the world of magazine publishing, that's not bad. Publishing companies know that profitability has to be measured across a portfolio of magazines, not unit by unit, and it's often worth your while publishing a loss-making Talk Magazine for a while for the touch of stardust glamour it adds to a lucrative (but potentially rather prosaic) Conde Nast Traveller. Slashdot would be a perfect "hood ornament" for a profitable stable of computer magazines, dragging the kids in while they were in college and then cross-promoting them onto other titles by the time they had reached a saleable demographic. And all this could be done without compromising its "editorial integrity", which is something usually respected in the media world, though not so much in the software publishing world ("Andover.net had all sorts of evil plans for Slashdot", our source reveals).

Bottom line: If Larry Augustin wants to claim to be running a company in the enterprise software business, it's time for him to walk the talk. Let's see some divestment of non-core assets like Slashdot. Otherwise, we ought to be facing facts and reminding ourselves that the company which used to be "VA Linux" and is now "VA Software" has always been "VA Media". It's a publishing company, and ought to be managed as one. If that means getting rid of Eleven Million Dollar Larry and getting a graduate of the Si Mewhouse academy, then so be it.