Deep Learning Startups May Encounter a Gotcha
March 14, 2020
Though fragmented, the deep learning AI market is growing rapidly. Anyone wishing to launch (or invest in) such a firm may want to check out Analytics India Magazine’s article, “Common Pitfalls that the Deep Learning Startups Fail to Recognise.” Writer Sameer Balanganur describes prevalent missteps under these headings: Not Investing Enough in Data and Powerful Processors, Not Accounting for the Cloud Charges, Expensive Data Cleansing, The Edge Cases, and Hiring the Right People.
The part that struck me was this description under Expensive Data Cleansing, as it Illustrates something many fail to understand:
“Training the model nowadays to achieve the state-of-the-art results [still] involves a lot of manual cleaning and labelling of large datasets. And the process of manual cleaning and labelling is expensive and is one of the largest barriers the deep learning startups face. … Although as time passes, the AI systems are moving towards complete automation, which will significantly reduce the cost. However, these AI-based automation applications still need human intervention for years to come. Even if there is full automation achieved, it’s not clear how much the margin of cost and efficiency will improve, so this becomes a matter of whether one should invest towards processes like drift learning and active learning to enhance the ability.
We noted:
“Not only expensive, the human intervention sometimes hinders the system’s creativity, but they might also do it by selecting what is essential for an algorithm to process or not using deep learning for a problem it can easily solve. Many times, deep learning is seen as overkill for many problems. The costs incurred by human intervention and cloud are interdependent. Reducing one means an increase in another.”
AI investment could be quite profitable, if one considers carefully. As always, look before you leap. See the write-up for more details.
Cynthia Murrell, March 14, 2020
Amazon Versus Microsoft: A Jedi Fight Development
March 13, 2020
DarkCyber spotted this story on the BBC Web site: “Pentagon to Reconsider Jedi $10bn Cloud Contract.” Since we are in rural Kentucky, the intrepid team does not know if the information in the Beeb’s write up is accurate. The factoids are definitely interesting. The story asserts:
The US Department of Defense is to “reconsider” its decision to award a multi-billion dollar cloud contract to Microsoft over Amazon.
The story points out that Microsoft is confident that its Azure system will prevail. Amazon, on the other hand, is allegedly pleased.
What’s at stake?
- Money
- A hunting license for other government contracts
- Implicit endorsement of either AWS or Azure
- Happy resellers, integrators, and consultants
- Ego (maybe?)
When will JEDI be resolved? Possibly in the summer of 2020.
Stephen E Arnold, March 13, 2020
Google Stadia: Google Wood or Just Recycled Cardboard?
March 12, 2020
DarkCyber does not play games. Sure, there are some young-at-heart DarkCyber games, but I ignore them. One of these hard-working individuals spotted “Google Stadia Hits an All-Time Low With This Embarrassing Tweet.” I am not much of a tweeter.
Apparently someone at Google does read tweets and noted one that contained this high school cheer / acrostic thing:
Note that there is no game for I.
A Googler replied, with a tweet, of course: “Why would you bring attention to this?”
I assume the answer is one of these choices:
a. It’s millennial or Gen X, Y, or Z humor
b. Stadia is not performing
c. Someone actually cares about Stadia to try to spell a word using the first letter of games on the service
d. There is a game on Stadia which uses the “what’s up” emoji instead of words.
The write up states:
Clearly, whoever is in charge of the Google Stadia Twitter account has stopped caring. It’s probably for the best since everyone else stopped caring about it months ago.
Google Stadia seemed doomed from the start, and things haven’t gotten much better. It lacks games, has a terrible monetization system, and generally isn’t all that convenient. It even pales in comparison to other similar systems like GeForce Now and Project xCloud. If the state of their social media is anything to go by, Google is already well on its way to just checking out and letting the system die. It’s hard to blame them. So far, Google Stadia seems like it was just a horrible idea.
DarkCyber has little insight to how things work at Google. I would surmise that whoever worked on Stadia has made an effort to catch on with a hot project team. No, not solving Death. Solving Stadia, however, may be a comparable challenge.
Stephen E Arnold, March 12, 2020
Amazon Versus Microsoft: JEDI in Play?
March 7, 2020
DarkCyber spotted a story in Stars and Stripes titled “Judge Says Amazon Likely to Succeed on Key Argument in Pentagon Cloud Lawsuit.” The source appears to be the Bezos-owned Washington Post. That fact may provide some context for the story.
The main point in the write up seems to be:
A federal judge has concluded that a bid protest lawsuit brought by Amazon over President Donald Trump’s intervention in an important Pentagon cloud computing contract “is likely to succeed on the merits” of one of its central arguments, according to a court document made public Friday [March 6, 2020].
The article states:
In an opinion explaining her reasoning, Campbell-Smith sided with Amazon’s contention that the Pentagon had made a mistake in how it evaluated prices for competing proposals from Amazon and Microsoft. She also concluded that the mistake is likely to materially harm Amazon, an important qualifier for government contract bid protests.
What’s missing from this story? Detail for one thing.
Several observations:
- Planners for the JEDI program are likely to experience uncertainty
- Regardless of the ultimate decision, time to implement newer systems is being lost
- The cost of the procurement process for JEDI will climb and, at some point, may become larger than the program itself.
Net net: Government procurement remains an interesting and impactful process. Procurement just keeps grinding its procedural mechanisms, delivering “efficiency.”
Stephen E Arnold, March 7, 2020
Honeywell: The Quantum Computing Thermostat Company
March 5, 2020
Yeah, that’s a bit of rural Kentucky humor. Honeywell is in four businesses and a fifth apparently has been added: Quantum computing. If you think Honeywell and recall the user friendly thermostat in your home, you are not thinking about the future, government contracts, breaking computing barriers, and putting technology pretenders like IBM, Google, and dozens of other companies in their place.
The Honeywell he CommercialPRO 7000 Programmable Thermostat is fantastic, according to Honeywell. For an entertaining experience, ask a friend to set the temperature for 4 pm today. This is a TikTok viral video DarkCyber believes.
To refresh your memory, DarkCyber wants to point out that Honeywell was once based in Wabash, Indiana. The firm generates about $40 billion a year from:
- Aerospace
- Building technologies
- Materials
- Safety productivity systems.
Now Honeywell is in the quantum computing business, according to the Wall Street Journal, March 4, 2020, edition. You may be able to locate the story behind a paywall at this link.
Honeywell has enjoyed a number of government contracts, and the firm is one of the leaders in smart controls and weapons management technology. In 1955, Honeywell teamed with Raytheon in order to compete with IBM. By the mid 1960s, Honeywell was one of the Snow White and the Seven Dwarfs of Computing. (Unfamiliar with this bit of digital history, Bing or Google may turn up some relevant hits, but I would recommend microfilm of the Minneapolis newspapers from this era. Don’t let your Bermuda shorts get in a bunch as you explore the innovations of Burroughs, Control Data Corp., GE, NCR, RCA, and my personal fave Univac.
Honeywell does a significant amount of computing and software/systems development. The firms owns a number of high technology business; for example, a radiation detection firm and has a stake in Zapata Computing.
Zapata says here:
We are the deepest bench of quantum scientists in the industry. Our founders helped create the field of near-term quantum algorithms including the invention of VQE, the progenitor of variational quantum algorithms.
The company’s approach relies on quantum charge coupled device (QCCD) architecture. The approach uses a technology called “trapped ions.” The idea is that useful work can be done due to leveraging mid circuit measurement. The idea is to insert a dynamic “if” based on the state of the calculation at a point in time. IonQ and Alpine Quantum Technologies also use the method. For some details, do a patent search for “trapped ion”. The background of US5793091A (assigned to IBM) provides some helpful information.
What business opportunities does Honeywell envision for its quantum computer? Here’s a selection gleaned from the PR blitz Honeywell launched a short time ago:
- Landing more customers like JPMorgan, Chase, and Company
- Speeding up financial calculations
- Creating new trading strategies (high speed trading?)
- Materials science applications (heat shields, stealth coatings?)
- Run Monte Carlo simulations (nuclear fuel analyses, risk and fraud analyses?)
The Honeywell quantum computer will be bigger than IBM’s quantum computer.
Interesting business play because Honeywell has a deal with Microsoft to plug the Honeywell technology into the Azure cloud.
The coverage of Honeywell’s announcement reveals the hyperbole associated with quantum computing. DarkCyber interprets the assertions as the equivalent of an athlete’s pre-season exercise routine. Progress may be made, but the effort can only be judged when the “star” is on the field and in the game.
Until then, the buzzword sells expectations, not a solution to a here-and-now problem. One has to admire Honeywell’s PR generating capability.
Stephen E Arnold, March 5, 2020
Import.io and Connotate: One Year Later
March 3, 2020
There has been an interesting shift in search and content processing. Import.io, founded in 2012, purchased Connotate. Before you ask, “Connotate what?”, let me say that Connotate was a content scraping and analysis firm. I paid some attention to Connotate when it acquired Fetch, an outfit with an honest-to-goodness Xoogler on its team. Fetch processed structure data and Connotate was mostly an unstructured data outfit. I asked a Connotate professional when the company would process Dark Web content, only to be told, “We can’t comment on that.” Secretive, right.
Connotate was founded in 2000 and required about $25 million in funding. The amount Import.io paid was not revealed in a source to which DarkCyber has access. Import.io, which has ingested about $38 million. DarkCyber assumes that the stakeholders are confident that 1 + 1 will equal 3 or more.
Import.io says:
We are funded by some of the greatest minds in technology.
The great minds include AME Cloud Ventures, Open Ocean, IP Group, and several others.
The company explains:
Starting from a simple web data extractor and evolving to an enterprise level solution for concurrently getting data that drives business, industry, and goodness.
What’s the company provide? The answer is Web data integration: Identify, extract, prepare, integrate, and consume content from a user-provided list of urls. To illustrate the depth of the company’s capabilities, Import.io defines “prepare” this way:
Integrate prepared data with a library of APIs to support seamless integration with internal business systems and workflows or deliver it to any data repository to develop robust data sets for advanced analytics capabilities.
The firm’s Web site makes it clear that it serves the online travel, retail, manufacturing, hedge fund, advisory services, data scientists, analysts, journalists, marketing and product, hospitality, and media producers. These are a mix of sectors and industries, and DarkCyber did not create the grammatically inconsistent listing.
Import.io offers videos which provide some information about one of its important innovations “interactive extractors.” The idea is to convert script editing to point-and-click choices.
The company is growing. About a year ago, Import.io said that it experienced record sales growth. The company provided a link to its Help Center, but a number of panels contained neither information nor links to content.
The company offers a free version and a premium version. Price quotes are provided by the company.
Like Amplyfi and maybe ServiceMaster, Import.io is a company providing search and content processing with a 21st century business positioning. A new buzzword is needed to convey what Import.io, Amplyfi, and Service Master are providing. DarkCyber believes that these companies are examples of where search and content processing has begun to coalesce.
The question is, “Is acquiring, indexing, and analyzing OSINT content a truck stop or a destination like Miami Beach?”
Worth monitoring the trajectory of the company.
Stephen E Arnold, March 3, 2020
Quantum Computing Dust Up: Is the Spirit of Jeffrey Influencing Some Academics?
March 2, 2020
If you are into quantum computing and the magic it will deliver… any minute now, you won’t bother reading the MIT Technology Review article “Inside the Race to Build the Best Quantum Computer on Earth.” Please, keep in mind that MIT allegedly accepted funds from the science loving Jeffrey Epstein and then seemed to forget about that money.
Here’s the key sentence in the write up:
None of these devices—or any other quantum computer in the world, except for Google’s Sycamore—has yet shown it can beat a classical machine at anything.
One minor point: MIT’s experts appear to have overlooked China, Israel, and Russia Is it really ignoring quantum computing?), to name three nation states with reasonably competent researchers.
The focus on IBM and Google is understandable. Did DarkCyber mention that IBM is contributing to MIT’s funding; for example, the IBM Watson Lab?
What’s the point of the MIT Magazine research? Let’s try to see if there are quantum-sized clues?
First, Google asserted in 2019 that the fun loving folks in Mountain View had achieved “quantum supremacy.” IBM responded, “Nope.” This write up expands on IBM’s viewpoint; specifically, Google’s quantum magic was meaningless. Okay, maybe from IBM’s point of view, but from Google’s, the announcement was super duper click bait.
Second, IBM is doing research and business development in parallel. Google sells ads; IBM sells … what? Consulting, mainframes, managed facilities, and Watson? Google sells ads. Ads generate money for Google moon shots and quantum PR. IBM spends its money on ads. Okay, that’s a heck of a point.
Third, IBM wants to build a quantum business that does business things. Google wants to build a cloud computer to [a] sell ads, [b] beat Amazon, IBM, and Microsoft in the cloud, [c] accomplish a goal like climbing a mountain, [d] it is just Googley, [e] two of the four choices.
Net net: The write up walks a fine line. On one side is IBM and its checkbook and on the other is the Google. Is the write up objective? From DarkCyber’s point of view, like artificial intelligence, quantum computing is just around the corner.
DarkCyber is checking to make sure that when NewEgg.com offers quantum components, the team can buy one. For now, we will stick with the Ryzen 3900x: It works, is stable, and does jobs without too much fiddling.
Quantum computers require a bit more work. But when deciding between funding and ads, maybe fancy dancing around quantum computing is the tune the MIT band is playing?
Stephen E Arnold, March 2, 2020
After Decades of Marketing Chaff, Data Silos Thrive
March 2, 2020
Here’s another round of data silo baloney—“Top 4 Ways to Eliminate Data Fragmentation Within Your Organization” from IT Brief. Surveys have found that many businesses are not making the most of all that data they’ve been collecting, and it has become common to blame data silos. It is true that some organizations could store and access their data more efficiently. There’s just one problem, and it is one we have mentioned before—there are some very good reasons to keep some data fragmented. Silos exist because of things like government requirements, legal processes, sensitive medical data, experts protecting their turf, and basic common sense.
The article asserts:
“Many organizations are finding it difficult to extract meaningful value from their data due to one endemic problem: mass data fragmentation. With mass data fragmentation, data volumes continue to rise exponentially, but companies struggle to manage that data because it’s scattered across locations and infrastructure silos, both in on-premises data centers and in the cloud. Organizations often don’t know what data exists, where it is and whether it’s being stored securely and in compliance with regulations.”
Of course, entities must ensure data is stored securely and that they comply with regulations. Also, the write-up’s advice to keep redundancies to a minimum and to understand how one’s data is being stored and accessed in the cloud are good ones. However, the exhortation to eliminate silos entirely is off the mark; trying to do so can be a fruitless exercise in expense and frustration.
Why?
- A person wants to hoard his or her information
- Rules or regulations prevent sharing to those “not in the fox hole”
- Lawyers and HR professionals don’t want legal documents available and “people” managers definitely do not want employee health and salary data flying around like particles motivated by Brownian motion.
Net net: Reality has silos. Accept it. Omit the marketing silliness.
Stephen E Arnold, March 2, 2020
Google: Feeling the Competitive Heat
February 28, 2020
Google, DarkCyber assumes, thought that Microsoft’s decision to convert the Chrome browser into Credge was a victory. “Google Is Now Warning Millions Of Microsoft Edge Users To Switch To Chrome: Here’s Why” tries to explain Googley thinking.
We learn from the capitalist tool:
Google has been found “abusing user agents,” the identifying code that enables websites to identify the browser type and version, to detect and warn Microsoft Edge users visiting the Chrome web store that when it comes to extensions they should switch to Chrome. The reason for the warning is that Microsoft Edge doesn’t integrate with the Safe Browsing protections Google uses to remove threats—so when an extension presents a risk, Google can’t act in the same way to protect users.
Is this the only reason?
DarkCyber thinks a bit of context will explain some of the Googley thinking.
Consider Google and Amazon.
Google does not like Amazon, especially when Amazon stepped away from solely being a retailer to offering software services to customers. Google wants some of Amazon’s cloud business, so they are telling retailers to chuck Amazon and check out their tools. ZDNet rolls out the gossip in the article, “Google’s Pitch To Retailers: We’ll Help You, From Search To Supply Chain.”
At the National Retail Federation, Google introduced retailers to a new line of tools available via its cloud. The tools range from product discovery, supply chain optimization, and hybrid application management. Thomas Kurian, Google Cloud CEO, explained that the retailers who innovate with their business plans are the most successful. Google wants to grab these forward thinking retailers with new tools like:
“Among the new offerings for retailers is a new tool called Google Cloud Search for Retail, which Google is piloting now and will introduce to the broader market throughout the year. The tool helps retailers improve search results for their own websites and mobile apps using cloud AI and Google Search algorithms.
Kurian’s blog post also served as a reminder to retailers that they can buy Google Ads to surface their products when customers use Google’s many consumer tools like Search, YouTube, Shopping, Google Assistant or Maps.
We noted:
Google also announced Google Cloud 1:1 Engagement for Retail, a set of best practices that can help retailers build data-driven strategies for personalized customer services. This should make it easier for customers to use Google’s BigQuery data analytics platform to build personalization and recommendation models.
That is just the beginning! Google is also developing a Buy Optimization and Demand Forecasting service that assisted retailers plan and manage supply changes. There is also a new retail version of Anthos, Google’s platform for managing services on site or the cloud environment. It will allow retailers to roll our and manage applications across all stores.
What happens if we try to add 1 + 1. DarkCyber thinks that the task reveals several facets of Googley thinking:
- Microsoft has lots of Windows 10 users who just use Credge. The browser works, is there, and why hunt for a different way to look at Web pages. But what if Credge gets traction on a mobile phone? What if Microsoft, the long time drone target of the Google, gets eyeballs on Android devices? Yep, those victory cheers are likely to become verbal and physical tics.
- Amazon is selling ads. Selling lots of ads is not good for the Google, a company for more than 20 years has had one revenue stream of significance. The Google wants to put some sand in the fuel tank of Bezos bulldozer. Thus, Googley behavior dictates action.
- Google itself faces a problem few companies have: Indexing the Web and changes to Web pages is expensive. How does one cut costs when Microsoft may blindly wreck havoc in the browser revenue flow or put a dent in the quite robust mobile ad business? How does Google protect existing ad revenue and possibly cause the Bezos bulldozer to go down for an engine overhaul? Aggressive actions seems to be the order of the day.
If DarkCyber steps back or in the lingo of a University of Chicago philosopher “go up a meta lever”, the actions of today’s Googlers reflect some changes which may give pause. Marketing has never been a Googley strength. Now it has to be competitive marketing. Is Google’s marketing elegant? Yeah, not too elegant.
Can Google control costs without further compromising its search service, its wonky innovations, and its increasingly contentious employee-management interactions?
DarkCyber finds the Credge and Bezos bulldozer “plays” interesting and entertaining.
Stephen E Arnold, February 27, 2020
Amazon Costs: Kubernetes to the Rescue
February 27, 2020
How much does an AWS customer pay for a specific service? DarkCyber knows that the taxi meter approach makes sense for the company that owns the medallions, the taxis, and possibly a few important people.
The taxi meter method in the cloud is a bit of a mystery — until the invoice arrives. “The Story Behind My Talk” explains a process which, according to Tuananh, can “reduce EC2 billing up to 80%.”
How does this money slip from the massive tractors of the Bezos bulldozer?
The article explains:
I started with a managed GKE at first using their free $300 credit, to learn the basic of Kubernetes; but then later on use kops to setup a production cluster with AWS.
Some of the changes I did for the production cluster is:
- Setup instance termination daemon to notify all the containers + graceful shutdown for all the apps.
- Setup multiple instance groups of various size and availability zone, mixing spot instances with reserved instances. This is to prevent price spike of certain spot instance group; and minimize the chances of all spot instances going down at the same time.
- Calculate and provision a slightly bigger fleet then what we actually need so that when there were instances shut off, there won’t be service downgrading. Because spot instances are so cheap, we can do this without worry much about the cost.
- Watch to see if there were scheduling failure to scale the reserved groups.
The payoff according to the article:
“The overall cost saving for EC2 was around 60-70% because we need to mix reserved instances in and provision a little higher than what we actually need. We were very happy with the result.”
Will AWS reconsider how it deals with EC2 billing? Does a bulldozer need diesel fuel?
Stephen E Arnold, February 27, 2020
-
- Subscribe to Beyond Search
Feature archive
News archive
-
