PODCAST
Gigaom Structure
by Gigaom Structure
A weekly podcast from GigaOM providing the latest news, analysis and interviews about the cloud, infrastructure and big data spaces.
-
5
The Hadoop wars, HP cloud(s) and IBM's big win
If you are confused about Hewlett-Packard's cloud plan of action, this week's guest will lay it out for you. Bill Hilf, the SVP of Helion product management, makes his second appearance on the show (does that make him our Alec Baldwin?) to talk about HP's many-clouds-one-management layer strategy. The lineup? Helion Eucalyptus for private clouds which need Amazon Web Services API compatibility; Helion OpenStack for the rest and HP Cloud Development Platform (aka Cloud Foundry) for platform as a service. Oh and there's HP Public Cloud which I will let him tell you about himself. But first Derrick Harris and I are all over IBM's purchase of AlchemyAPI, the cool deep learning startup that does stuff like identifying celebs and wanna-be celebs from their photos. It's a win for [company]IBM [/company]because all that coolness will be sucked into Watson and expand the API set Watson can parlay for more useful work. (I mean, winning Jeopardy is not really a business model, as IBM Watson exec Mike Rhodin himself has pointed out.) At first glance it might seem that a system that can tell the difference between Will Ferrell and Chad Smith might be similarly narrow, but after consideration you can see how that fine-grained, self-teaching technology could find broader uses. AlchemyAPI CEO Elliot Turner and IBM Watson sales chief Stephen Gold shared the stage at Structure Data last year. Who knows what deals might be spawned at this year's event? Celebrity_ChadSmith_WillFerrell_cropped Also we're happy to follow the escalating smack talk in the Hadoop arena as Cloudera CEO Tom Reilly this week declared victory over the new Hortonworks-IBM-Pivotal-backed Open Data Platform effort which we're now fondly referring to as the ABC or "Anyone But Cloudera" alliance. It's a lively show so have a listen and (hopefully) enjoy.
-
4
Mark Cuban on net neutrality: FCC can't protect competition
As someone who takes her cues on net neutrality from Gigaom's resident expert Stacey Higginbotham or, failing that, John Oliver, this is hard to admit: Mark Cuban may have a point on why the proposed net neutrality regulations may be a cure that's worse than the disease. If adopted, he maintained, these regs will open the door to more confusion, more litigation and more overall turmoil, none of which will serve consumers well. Before you throw your device at the wall, just give him a listen. Cuban, the serial entrepreneur who started out as a VAR before founding Broadcast.com which sold to [company]Yahoo[/company] in a $5.7 billion stock deal in 1999. He is now owner of the Dallas Mavericks, co-star of Shark Tank and CEO of of AXS TV and interestingly a star of new AT&T commercials. The hyphens just keep coming. He sees competition ramping up in both in wired and wireless access -- if these markets are so foreclosed why is Google doing broadband? Why is "AT&T going out of its traditional TV markets where they have U-verse to compete with Comcast and Google? That's one layer. On the other layer you have mobile, with [company]Cablevision[/company] going into Manhattan where [company]Verizon[/company] and AT&T have broadband wireless and putting together an unwired wifi network for $30 a month' His point is that there is competition, although it may not be the competition we would all like to see. Cuban is clearly worried about one, well two mega players and neither one is a big ISP. "I would rather see national competition for [company]Google[/company] than no competition for Google. If you put a lid on Time Warner and Comcast and Google just keeps adding more and more markets, who's going to compete with them?" Google and [company]Apple[/company] constitute a huge countervailing force for all the ISPs because of their mobile might. "The fastest growing access for Internet is mobile. Who controls access to mobile? Google and Apple. The far greater risk is if Apple decides that the Comcast app is not right, Comcast won't be able to reach most of its market to give access to its own broadband. Kind of crazy but it's a possibility." For the record, he isn't recommending regulation to stop that either. His point isn't that Comcast or Time Warmer or insert-your-least-favorite cable provider here) are so great -- he admits they are not -- it's just that the FCC its regulations are ill equipped to deal with fast-changing technologies. The public would be better served to let the cable companies duke it out with each other and, perhaps more to the point, with far scarier competitors including Google and Apple. He starts about 10 minutes in. But here is the kill shot: Do you really want the same organization (the FCC) that took 8 years to deal with Janet Jackson's Wardrobe Malfunction at Super Bowl XXXVIII to be the gating factor in the internet? Ummmm, maybe not. Listen to the whole thing to find out how you, too, can get in touch with Cuban, such a shy and reserved guy, to ask your own questions on net neutrality; whether the NBA is seeing diminishing returns on data analytics; and why the heck the Celtics let the Mavs steal Rajon Rondo. Whatever. Businessman and TV personality Mark Cuban speaks onstage at TechCrunch Disrupt at Pier 48 on September 8, 2014 in San Francisco, California. Businessman and TV personality Mark Cuban speaks onstage at TechCrunch Disrupt at Pier 48 on September 8, 2014 in San Francisco, California. In our intro section, Jonathan Vanian and I discuss all (or a bunch anyway) of this week's Kubernetes news -- where Mirantis was latest into the pool, working with Google to bring the cluster management framework to OpenStack clouds, joining HP and a raft of other tech vendors endorsing the open-source framework. Interestingly, Spotify blazed its own trail, Helios as opposed to Kubernetes for its own workloads.
-
3
Microsoft's machine learning guru on why data matters sooooo much
Not surprisingly, Joseph Sirosh, has big ambitions for his product portfolio at Microsoft which includes Azure ML, HDInsight and other tools. Chief among them is making it easy for mere mortals to consume these data services from the applications they're familiar with. Take Excel for example. If a financial analyst can, with a few clicks, send data to a forecast service in the cloud, then get the numbers back, visualized on the same spreadsheet, that's a pretty powerful story, said Sirosh who is corporate VP of machine learning for Microsoft. But as valuable as those applications and services are, more and more of the value to be derived from computation over time will be the data itself, not all those tech underpinnings. "In the future a huge part of the value generated from computing will come from the data as opposed to storage and operating systems and basic infrastructure," he noted on this week's podcast. WHich is why one topic under discussion at next month's Structure Data show will be who owns all the data flowing betwixt and betweeen various systems, the internet of things etc. When it comes to getting corporations running these new systems Microsoft may have an ace in the hole because so many of them already use key Microsoft tools -- Active Directory, SQL Server, Excel. That gives them a pretty good on-ramp to Microsoft Azure and its resident services. Sirosh makes a compelling case and we'll talk to him more on stage at Structure Data next month in New York City. In the first half of the show, Derrick Harris and I talk about the Hadoop world has returned to its feisty and oh so interesting roots. When Pivotal announced its plan to offload support of Hadoop to Hortonworks and work with that company along with IBM, GE on the Open Data Platform the response from Cloudera CEO Mike Olsen in a blog post with his take. Also on the docket, @WalmartLabs massive OpenStack production private cloud implementation.
-
2
No, you don't need a ton of data to do deep learning
There are a couple of seemingly contradictory memes rolling around the deep learning field. One is that you need a truly epic amount of data to do interesting work. The other is that in many subject areas there is a ton of data but it's not just laying around for data scientists to snarf up. On this week's Structure Show podcast, Enlitic Founder and CEO Jeremy Howard and Senior Data Scientist Ahna Girshick address those topics and more. Girshick, who is our first guest who's worked with Philip Glass and Björk on creating music visualizations, said while there may be scads of MRIs, CAT scans, x-rays created -- they're typically used for their primary purpose -- to diagnose your bum knee, and are then squirreled away in some PACS system never to see the light of day again. All of that data is useful for machine learning algorithms, or would be, if it were accessible, she said. Girshick and Howard agreed that while deep learning -- the process of a computer teaching itself how to solve a problem -- gets better with more data, there's no reason to hold off working with it for data become available. "While more data can be better I think this is stopping people from trying to use big data inappropriately," Howard said. He cited a recent Kaggle competition on facial key point recognition that's using 7,000 images and "the top algorithms are nearly perfectly accurate." The reason companies like Baidu and Google say you need mountains of data is because they have mountains of data available, he said. "I don't think people should be put off trying to use deep learing just because they don't have a lot of data." Enlitic is using deep learning to provide medical diagnoses faster and help provide better medical and outcomes for millions of underserved people. It's a fascinating discussion so please check it out -- Girshick will speak more on what Enlitic is doing at Structure Data next month. And, if you want to hear what's going on with Pivotal's big data portfolio, Derrick Harris has the latest. Oh and Microsoft makes a bold play for startups by ponying up $500K in Azure cloud credits starting with the Y Combinator Winter 2015 class. That ups the ante pretty significantly compared to what Amazon Web Services, Google and IBM offer. Your move boys.
-
1
VMware wants all those cloud workloads "marooned" in AWS
Even though VMware initially called its Amazon competitor vCloud Hybrid Services, make no mistake, it's the company's public cloud (now renamed vCloud Air.) And, [company]VMware[/company] really wants workloads that might run ow on [company]Amazon[/company] Web Services to come on over, says Bill Fathers, EVP and GM of cloud services for VMware. That's a tall order. Face it, AWS has been around as, an old boss would have said "Hector was a pup" The first services launched in 2006, and vCloud Air is, what?? a old toddler of two. Fathers acknowledged vCloud Air doesn't have thousands of customers but said that wasn't the plan. Instead VMware wants a few hundred of key companies to act as bellwethers. But Fathers point is that a small percentage of total computing is now running on any public cloud -- he thinks it's now 5 or 10 percent up from the two to three percent he thought it was last June at Structure. Which means that there's a ton of work up for grabs. And while AWS looks to be the enemy for VMware's cloud, the same is not true for Google -- VMware last week announced plans to offer and support four Google services including BigQuery, on vCloud Air. This week it brought out its promised Integrated OpenStack. Fathers positions both the Google relationship and this week's Partner Exchange announcements. And he's clearly not backing away from a fight with the biggest of big clouds. Bill Fathers, VMware EVP and GM cloud services Beyond the VMware universe there was a bit of big data moving and shaking with Cloudera buying Explain.io and its self-service query modeling expertise and Datastax picking up Aurelius, the keeper of the Titan graph database. Could this be a sign of even more M&A to come? That's something we'll hear more about at Structure Data from March 18-19 in New York from the CEOs of Cloudera, Hortonworks and other data powerhouses, so book your tickets now. Ok, the gratuitous plug is now done, it's time to listen.
We're indexing this podcast's transcripts for the first time — this can take a minute or two. We'll show results as soon as they're ready.
No matches for "" in this podcast's transcripts.
No topics indexed yet for this podcast.
Loading reviews...
ABOUT THIS SHOW
A weekly podcast from GigaOM providing the latest news, analysis and interviews about the cloud, infrastructure and big data spaces.
HOSTED BY
Gigaom Structure
Loading similar podcasts...