Since 2000, company has been based in Tokyo and through its social learning service iKnow has focused on helping Japanese speakers learn English. But on Tuesday, the company announced that it has opened a stateside office in Palo Alto, Calif. and is launching a new product to help people learn and remember any kind of content.  The web-based tool is currently in private beta (the first 500 GigaOM readers can click here for an invite) but is expected to be launched more widely in early 2013.

While several new startups and learning platforms provide formal students and lifelong learners the opportunity to take courses on all kinds of subjects, Andrew Smith Lewis, Cerego’s cofounder and executive chairman, says his product applies learning principles drawn from neuroscience and cognitive science.

As students progress through courses on the site, Cerego takes a “spaced rehearsal” approach, which supports a learning technique that involves the repetition of content over increasing periods of time, to calculate the optimal moments to review content. The algorithms consider what students got right and wrong, as well as their familiarity with related content and, potentially, what others on the platform found challenging or easy to determine how likely they are to forget specific content items and when that content should be reviewed.

“They’re like interactive notecards that are smart and know exactly what you know and don’t know,” said Lewis.

The site has been seeded with about 50 courses on topics from exotic animals to statistics to American cuts of beef, but the goal is for users – whether they’re students, professors, casual learners, publishers or even corporations – to add to the site with their own content. College students could use it to study for a test on anatomy or adult learners – including those taking courses on online learning sites like Coursera and Udacity – could use it to review programming terminology, Lewis said.

The company said the basic service will remain free, but it plans to charge for premium content from validated publishers as well as premium services, such as advanced analytics or functions for power users.

To date, Lewis said Cerego has raised $28 million from private investors and claims companies including SoftBank and Yahoo Japan as clients. It’s also attracted an interesting group of advisors, including Scott McNealy, co-founder Sun Microsystems and Joi Ito, director of the MIT Media Lab. Since launching, 1.1 million Japanese users have logged 3.7 hours on the platform, and Cerego says all of that data has been used t refine the algorithms for its latest product.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• 12 tech leaders’ resolutions for 2012
• Themes for a connected world: GigaOM RoadMap review
• The future of mobile: a segment analysis by GigaOM Pro

I confess that when I first saw my friend Michael’s face appear in my chat list, I was taken aback — and more than a little disturbed by it. It was a couple of weeks after his funeral, and so the memory of his death had faded to some extent, and his smiling picture felt like a rude reminder. It reminded me of web articles I had seen about how (or whether) to delete deceased friends or family members from Facebook’s social graph, and at first I thought about doing that.

But then I thought about how difficult it had been deleting another friend’s contact information from my cellphone after he died (this was before Facebook had become popular) and how it felt like I was deliberately forgetting about that person, which didn’t feel right.

How social do we want death to be?

It occurred to me that we often keep photos of loved ones in our wallets or in picture frames on our mantelpieces, as a way of remembering them after they are gone. I have pictures of my father, who died more than a decade ago now, as part of a random photo slideshow that comes up on a spare computer and on the television for the same reason. So why does it feel so different when we see that person’s avatar pop up in our Facebook feed or a chat window? Perhaps because social media is inherently about communication — and in most cases real-time communication — and that person can no longer be communicated with.

Facebook has a process whereby a person’s page can be “memorialized,” or turned into a kind of static page as a tribute to them, where friends and loved ones can post and see messages posted by others but access is restricted and it doesn’t show up in recommended lists (you can ask the social network to do this by filling out a form). In many cases, particularly when young users die in some violent or tragic way, their friends turn the page into a memorial quite quickly — and of course journalists then often show up asking for comments or photos, which brings up a host of other questions about what’s appropriate.

But if the page belongs to someone who hasn’t really been a public figure, and didn’t die in any kind of newsworthy way, then it falls into a kind of grey area. Do you maintain the page? Mothball it? Eventually delete it? In the case of my friend Michael, who was a fairly prominent user of social media in his job as a marketing professional in Toronto (one of the reasons I believed he wouldn’t mind my live-tweeting his funeral), his family chose to keep the page alive — and has even posted messages to him as though he was still around, which I find heart-warming in an odd way.

When real death meets virtual death

And Facebook is just one part of the equation when it comes to handling a person’s social media after they die. What about their Twitter account, or their Tumblr account, or even their email? When my father-in-law died, the family was confronted with a dilemma because he and his wife had shared an account that used both of their names — so when an email came in from my mother-in-law, his name showed up in the address field as well, which was somewhat uncomfortable. But changing email addresses is not easy.

There are also issues around who owns a user’s social content after they die: does Facebook own that person’s page and status updates and photos, and if so what duty do they have to provide it to family members? What about iTunes? Twitter is less of an issue because no one can get access to their tweets anyway, even if they are alive (unless they make a special request, as Andy Carvin of NPR did for his tweets during the Arab Spring). But what about Flickr photos or Pinterest pages? It’s still a somewhat unexplored region of our online lives at this point.

But for me, the more interesting aspect is how we look at all of those pages and tweets and photos and avatars. Are they a welcome reminder of that person and how we used to fit into their lives, or are they a cruel joke played on the living because they seem to promise a level of interaction that we will never be able to have again? Perhaps they are both — and perhaps it is too much to ask that our virtual worlds be any more comfortable around death than our offline ones are.

Post and thumbnail images courtesy of Flickr user Chad McDonald

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Facebook’s IPO filing: ideas and implications
• NewNet Q2: Google closes the quarter with a bang
• What Enterprise Software Vendors Could Learn from the Consumer Space

If you’re a tech company trying to speed up applications and provider faster access to more memory in your servers, the technology could be your holy grail. Everspin, the Chandler, Ariz,-based startup pioneering MRAM, said today that its spin-torque MRAM that it hopes to use to replace DRAM, is sampling and would be available for use in products by 2013. MRAM is one of several new evolutions in memory technology designed to make applications perform faster by removing bottlenecks in getting the data from storage to the processor. Today, even flash-based systems can still cause a delay in application performance.

Steffen Hellmold, VP of marketing of Everspin, says that MRAM in its current version is about 100 times faster than flash, with the caveat being that he’s comparing MRAM to basic flash, while most companies tweak their flash arrays to make them faster and more reliable. Another caveat is that MRAM costs about 10 times more than flash-based solutions, although for certain applications trading 100 times the performance at 10 times the price is an easy call to make. The final caveat is that the current samples are for 64 Mb memory sizes which is less memory than your home PC had in the mid 90s. Other types of memory in this emerging niche include phase-change memory and memristors, which are pretty far away from commercialization.

For more on MRAM check out this article from early this year, or a background[ER] on Everspin written when it was spun out from Freescale Technology back in 2008.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Report: How Mobile Cloud Computing Will Change Tech
• The importance of putting the U and I in visualization
• AWS Storage Gateway jolts cloud-storage ecosystem

Here’s how the company explains the technology in the press release announcing it:

The Extended Memory subsystem dynamically moves frequently accessed data pages into memory on-demand while transparently migrating rarely accessed data pages from DRAM into ioMemory. This allows developers to simplify application design by assuming that entire datasets are in-memory, without the costs associated with DRAM purchase and operation. Application developers are able to further tune performance through software development kit tools that lock selected pages into DRAM, giving access to NAND flash as memory, instead of treating it as an extension of disk storage.

Fusion-io was able to pull this off in part because of an industry trend toward higher-capacity, less-expensive flash storage and in part because Fusion-io doesn’t claim to offer solid-state drives. When I covered the release of the company’s ioDrive2 and ioDrive Duo2 (which plug into a server’s PCI slot and operate as repositories for data needing high performance) in October, the products represented a more than 2x capacity gain over the previous generation and continued to boost IOPS performance, despite the smaller and cheaper NAND technologies on which they’re built.

Fusion-io Founder and CEO David Flynn attributed his company’s ability to continually raise performance on new, consumer-driven NAND designs with its decision to to provide memory controller software rather than trying to act like a hard-disk drive (as other solid-state drive providers try to do) and use a microcontroller. That same decision, the company says, also helped make Extended Memory possible because it didn’t have to try and make disk-storage protocols function at the memory tier.

In-memory databases and analytic systems (such as SAP’s high-performance HANA appliance) are nothing new, but they’re limited in size to the system’s memory footprint and can be rather expensive. All-flash storage arrays are also becoming much more popular and getting less expensive, but they still exist as a separate storage tier and are often aimed at legacy enterprise applications. If what Fusion-io and Princeton have created works as promised, it could create a middle ground approach that resonates with customers — such as Fusion-io mega-user Facebook — that need a workable combination of massive scalability and high performance.

Image courtesy of Shutterstock user Mark Schwettmann.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Infrastructure Q1: Cloud and big data woo enterprises
• A near-term outlook for big data
• Infrastructure Q4: Big data gets bigger and SaaS startups shine

London-based Memrise has a cool take on learning and memorization, combining vivid encoding techniques such as visual tools and mnemonic devices with the kind of game mechanics you might find in a Zynga game. The company can take a word in a foreign language and create an animated cartoon or a “mem” around it, giving a user a visual way to remember that particular word. It then schedules reminders and tests of mems to ensure they stick in a user’s long-term memory.

The goal is to move a new word from a greenhouse for short term memory into a long-term memory garden. Memrise will take into account the difficulty of particular words and tailor tests for users that ensure it doesn’t die. Ultimately, it’s supposed to make learning fun, like a recreational activity.

“We found game dynamics are useful to allow people to relate emotionally to memories. That’s not easy because memories are organic beasts that fade,” said memory grandmaster Ed Cooke, who founded Memrise with Greg Detre, a Ph.D. in neuroscience from Princeton.

Memrise fully supports six languages: French, Spanish, German, SAT Vocabulary, Mandarin and Italian, with more than 2,000 full audio and mnemonics-enabled mems. But its user community has added another 1 million words. Memrise is looking to expand beyond languages to anything factual and recently began a partnership with the Guardian newspaper in the UK, teaching readers about things like cheeses, herbs, plants and animals.

Cooke said the company is poised to release mobile apps in the next few months. Memrise is looking at making money through a premium model for the mobile apps, which could have extra features available through a subscription, he said. Further down the road, the company is looking to relocate to San Francisco or New York.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• How emerging technologies will influence collaboration
• Facebook’s IPO filing: ideas and implications
• Defining work in the digital age: an analysis by GigaOM Pro

To understand how cool this is, you have to understand the problem.

The big power problem

The chip industry is really good at making a CPU that does calculations faster, but it hasn’t been able to make memory chips fast and dense enough to feed the cores enough information to keep up with the CPU’s capabilities. So chips are left with is a massively large brain that stands idle sometimes while it waits for information to come to it. That idle time burns power and reduces the overall performance of a computer — and it’s becoming a bigger deal as both power and performance are being pushed to the edge.

Many companies such as Microsoft, Intel, and even startups such as Tilera are looking for ways to solve the what the industry calls the memory bandwidth problem. Samsung and Micron’s contribution with Hybrid Cube Memory is to deliver a chip that can send 15 times more data than a common DDR3 DRAM module used today. The consortium claims in a presentation that its new technology will use 70 percent less energy per bit than existing DDR3 DRAM technologies.

A dramatically different architecture

Some companies try to address the memory bandwidth problem by creating custom fabrics inside the chip to shuttle information around, while others basically create a giant caching system to pull information as rapidly as possible. The Hybrid Memory Cube guys are approaching it with density and a logic layer in a different type of architecture. The logic layer sits on the bottom, and the memory is densely stacked on top in a cube as opposed to a flatter architecture. Stacked memory isn’t new either, as startups and large chip firms like IBM  have tried that approach — but such a radically new architecture as a densely stacked cube has its pros and cons.

Rethinking the way memory chips are built allows for massive improvements in performance. That, in turn, will help drive faster supercomputers and help computers take full advantage of the coming improved speeds in broadband network bandwidth (pumping a gigabit connection to server doesn’t help if the processor inside can’t pull information in for processing at gigabit speeds). It will also help support the gear that is needed to deliver that bandwidth.

The chip is just the beginning

But, the chip industry is the first layer in a huge stack of hardware and applications, which means architecture changes are hard to push without the consent of the folks on top of the stack. A chip’s technology may be the most awesome thing ever, but to get it out of the niche market, big-time equipment makers and software firms will have to demand the technology. A perfect case for this might be Infiniband, the networking technology that was supposed to take on Ethernet, but instead ended up as an awesomely fast networking technology relegated to the high performance market.

Luckily, the Hybrid Memory Cube guys are solving a huge problem that’s been a pain point for the industry for a few years, and will only get worse. It’s also starting in a niche market where the technology can be used first, become more widely adopted, and eventually put into production in a manner that might let it gain some economies of scale and drive down the cost of the chips. It’s also trying to solicit folks across several industries to join the effort and embrace the technology. Altera Corporation, Open Silicon, and Xilinx  are working with the consortium to define a specification to enable applications ranging from large-scale networking to industrial products and high-performance computing to run on the chips. That specification should be out in 2012. Let’s see if this licks the memory bandwidth problem, and whether or not the industry elects to embrace it.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• The future of mobile: a segment analysis by GigaOM Pro
• GigaOM Research highs and lows from CES 2013
• The state of cross-platform media measurement

Science magazine has published some research into how our memories are influenced by the availability of computers as a source of information, and this has some in a tizzy about the implications of outsourcing our brains. Author Nick Carr, for example — who has written a whole book about how the web is changing the way we think and making us more shallow — says he worries this phenomenon is going to make us less human in some way. But is that really a risk? I don’t think so. I, for one, am glad to outsource the duty of remembering miscellaneous facts to the cloud, because it leaves me free to do more important things.

In a nutshell, the Columbia University psychologists who published the study performed a number of experiments designed to test whether subjects remembered certain things better or worse when they were told that the information — such as “An ostrich’s eye is bigger than its brain” — would be stored in a computer somewhere or would be available through a search engine. Not surprisingly perhaps, people’s memories were somewhat less reliable when they knew the answers they were seeking would be stored for later retrieval (there are more details at the Columbia website).

Implanting forgetfulness?

Carr says he’s worried that by losing these facts and details we store elsewhere, we will become less human in some way, or lose some core of ourselves. But is that really what’s happening? I don’t think so. It’s not like I’m suddenly going to forget my son’s first steps (oh, that’s right — I have daughters!) because I use Google to look up who starred in that movie we watched a couple of years ago, or to figure out who the head of the United Nations is. It’s worth remembering that the invention of writing triggered similar fears, as Plato reminds us in The Phaedrus, quoting the King of Thebes:

If men learn this [writing], it will implant forgetfulness in their souls;they will cease to exercise memory because they rely on that which is written, calling things to remembrance no longer from within themselves, but by means of external marks. What you have discovered is a recipe not for memory, but for reminder.

Carr also makes the argument in his book “The Shallows: What the Internet is Doing to Our Brains” that we are becoming not just dumber as a result of the web, but also (supposedly) less interesting, because our brains are being trained to focus on the ephemeral and the trivial instead of the important things we should be spending time on. I took issue with this kind of fear at the time, as did some others, and I think Carr is being similarly alarmist in this case. Besides, if we use the cloud to remember the trivial and ephemeral for us, wouldn’t that be a good thing by Carr’s definition?

Do we still need to memorize things?

I know that in my parents’ time, memorization of huge lists of facts and figures and Shakespearean sonnets was standard, because that was the criteria by which knowledge was judged. But what difference does it really make if I can’t remember when the War of 1812 was? (that’s a joke, by the way). Is my experience of the things that matter in life going to be impaired because I don’t know who signed the Magna Carta? I can see how this would be a problem if a trivia game suddenly comes up while I am camping in the woods, but other than that, I don’t see why I shouldn’t outsource that to the cloud — the same way lots of people used to outsource it to Encyclopedia Britannica.

As one commenter on Google+ mentioned when I shared the Science magazine article in my stream, the benefit of having something like the Internet available at all times is that it is the most comprehensive collection of knowledge ever invented (although obviously not all of it is correct). How can that not be a good thing? Said Justin Fogarty:

The plus side is that the whole of human knowledge is nearly at our fingertips. I will not miss card catalogs, the Dewey decimal system or heavy book bags.

Computers can’t really replicate memory anyway. All they can do (so far, at least) is store facts — but facts are not memories. What real memories are made up of is smells and sounds and emotions, and no computer or cloud-based system can store those things. But what the cloud can do quite well is store my phone numbers and the photos I took on a particular day or the tweets I sent (something an app called Momento is extremely good at) and leave me free to relive the memories associated with those facts.

To me, that’s a fair trade — the cloud remembers all the boring and mundane details and facts of my life (yes, I use Facebook to remember when people’s birthdays are, as I expect a lot of people do) and I get to focus on the things that are really important.

Post and thumbnail photos courtesy of Flickr users Stefan and Tim O’Brien

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• A clouded view of Google Music
• What Enterprise Software Vendors Could Learn from the Consumer Space
• The evolution of consumer-media cloud storage

According to Storm On Demand, its new behemoth cloud server features 2.0 GHz per core, reads at 3.3 Gigabits per second and writes at 4 Gigabits per second. A company spokesperson told me via email that customers doing image rendering have been particularly excited about the new instance size, and that the National Republican Senatorial Committee launched an instance just today to power a very large database.

That sounds about right. Lots of memory is ideal for situations where serving the data from disk becomes a bottleneck because the hard drive can’t keep up with today’s high-powered processors.

Perhaps the best thing about the 96GB instance is that it costs only $1.37 per hour. That’s less than the cost of the largest servers available from other cloud computing providers, and they aren’t nearly as big. The most-comparable option comes from Amazon Web Services, whose High-Memory Quadruple Extra Large Instance comes with 68.4 GB of memory and 26 EC2 Compute Units (8 virtual cores with 3.25 EC2 Compute Units each). But it costs at least $2 per hour.

Storm On Demand is establishing quite a reputation for delivering high-performance cloud servers. As I highlighted in a recent GigaOM Pro piece (sub req’d), cloud-benchmarking service CloudHarmony found that Storm On Demand consistently ranked the best, or among the best, across its gamut of tests. In terms of memory I/O, Storm On Demand’s previous top instance — coming in at 48GB and 12 cores — outperformed every other major cloud provider’s fastest option. The only two that came close were AWS’s Cluster Compute Instance, as well as its aforementioned High-Memory Quadruple Extra Large Instance.

As I also highlight in my GigaOM Pro post, though, cloud benchmarks are only a starting point because performance will always vary for different users and different applications. Even though Storm On Demand looks like the real deal in terms of cloud computing performance, anyone serious about achieving a certain result for their applications really should pay the nominal amount to test a number of options.

Image courtesy of Flickr user richardmasoner.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Benchmarking the Cloud: Your Mileage May Vary
• Infrastructure Q3: OpenStack and flash step into the spotlight
• The fourth quarter of 2012 in cloud

RAM

This is the easiest place you can make an improvement with almost any Mac. iMacs have a user-accessible memory slot located on the bottom edge of the display, between the built-in speakers. It’s as simple as removing three screws, and then replacing the computer’s existing memory with large capacity, compatible RAM units. MacBooks and MacBook Pros are also easily upgradeable when it comes to memory, and require only the removal of the bottom case or a memory area cover, depending on your model.

Remember to make absolutely sure that the RAM you’re buying is compatible with your computer. You can do this by finding out which Mac you have, by doing the following:

1. Go to the Apple menu in the top left corner of the menu bar.
2. Click on “About This Mac.”
3. Click on the “More Info…” button.
4. Take note of the Model Identifier in the Hardware Overview that opens. It should say something like “iMac12,2″ or “MacBook3,1.”
5. Go to EveryMac.com and enter that identifier to get the complete specs for your machine, including what type of RAM it uses and the maximum amount it can support.

If you’d rather make it easier on yourself, you can go to OWC and browse for your model in the Memory section of their online store. Note that this does require that you at least know when you bought the computer, but you can find that out using the method described above.

Note that some Mac mini models and MacBook Airs make upgrading the RAM yourself very difficult or impossible, in which case you might want to consult with an authorized Mac service provider.

Hard Drive

RAM is easier to upgrade, but for Lion, a hard drive change could make the biggest difference. They can be a bit trickier to upgrade, and nearly impossible if you have an iMac, but for most Mac notebooks, the process isn’t very challenging. If you’re unsure how to change your hard drive, check the official Apple manual for your computer, which you can identify using the method described above.

There are a few options available to you here. Like Dave Greenbaum, you could choose to go with a solid-state drive. This is the best possible option, but it’s also the priciest, and you get relatively little storage space for your money.

Another option is to supplement your existing HDD by installing a separate, smaller capacity SSD as your startup volume. If you have a MacBook Pro, this isn’t too difficult to do. Weldon Dodd provided an excellent walkthrough of how he achieved this with his own computer. Lion requires a minimum of 8 GB of free space on your drive, but a 40 GB drive that you can get for around $100 is probably as small as you should go, even if you don’t plan on keeping anything else on there.

Finally, you can also get the best of both worlds in a single drive, using a hybrid drive. Seagate makes a hybrid drive that provides 500 GB of storage, along with a 4 GB flash storage module. It manages to achieve a 32 MB cache with a 7200 RPM drive speed, and provides much better performance than a traditional drive, in my experience, plus it’s way cheaper than a standalone SSD. Boot times, app launch times, and wake from sleep times all drastically improved after I installed this in my MacBook Pro.

The most expensive of the above-mentioned options probably won’t cost you more than $500, and that’s only if you opt for a relatively capacious SSD. Noticeable improvements shouldn’t cost you more than $100, which is a great deal shy of the $1000+ you’d spend on a brand new Mac. Any other suggestions for DIY improvements that could make transitioning to Lion that much more satisfying?

* Ed. Please remember that DIY modifications to Apple equipment is done at your own risk, and voids your AppleCare warranty.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Research In Motion: future scenarios for its fate
• The evolution of consumer-media cloud storage
• Who and what to watch in the new era of the living room

Kaminario sells the K2 appliance, which delivers faster access to more information via a blade-based DRAM array. The DRAM array has cost advantages over Flash-based solid-state drives, which puts Kaminario in a slightly different category than the other SSD folks who are generally using Flash memory. However it’s still banking on similar trends of enterprises needing faster access to more data. Kaminario launched last June about the same time as Anobit, another storage vendor that is offering a cheaper variation of SSDs (Anobit shares Pintango as an investor).

Joining their ranks are Violin Memory, Pliant and Fusion-io, which filed in March for an initial public offering. At the time I said, and still believe, that it would much rather get bought. So, with all these startups trying to sell some version of better, faster, cheaper SSDs into the enterprise, who’s likely to pick up the technology? Obvious buyers would be Intel, which is a huge vendor of SSDs, any of the memory giants Samsung, SanDisk, Seagate or Western Digital, or perhaps even a computing vendor such as Dell Hewlett Packard or IBM if the server becomes just another component in the data center.

Because Violin and Kaminario both sell storage appliances they may be a better fit for storage vendors who get a complete product line, whereas Pliant and Fusion-io are more in the component space and may be a better product for storage or server vendors. Anobit, because it offers a way to boost the performance of a cheaper type of Flash memory could be a good fit with an actual memory maker such as Samsung or Intel.

Image courtesy of Wikipedia Commons contributor Greg L. Wright.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Infrastructure Q2: Big data and PaaS gain more momentum
• In Q3, Big Data Meant Big Dollars
• Infrastructure Q3: OpenStack and flash step into the spotlight