In this issue of Cloudnomics: The Ultimate Guide to Google’s Speech AI… Google’s new Cloud Fleet Routing API… Using Bigtable to handle 5M+ ad requests per second… Plus how GCP’s Security-by-Design keeps you safer…
How To Speak Google
Whether you’ve recently chatted with your smartphone and asked it to tell you the location of the nearest Chipotle, or you’ve overheard your 2-year-old bossing “Awexa!” around, you’re probably aware that we’re using speech more and more as a way to converse with our machines.
(Courtesy of Google Cloud's 90's Action Movie Catchphrase API - still in beta)
In this Ultimate Guide to Speech on Google Cloud, Dale Markowitz, an Applied AI Engineer at Google, describes several ways you can use Google’s Cloud Speech API, including captioning, real-time translation, sentiment analysis, and creating voice-driven interfaces (ie: interactive virtual agents, smart speaker and smartwatch apps, etc.).
The best part? Once you’ve mastered Google’s Cloud Speech API, the next time a buff robot from the future with an Austrian accent tells you, “Hasta la vista, baby!” you’ll be able to translate, transcribe and caption his message in real time as you run away for your life!
Bonus: Dale also made this nifty tutorial on translating and dubbing videos with AI, using Google Cloud’s Speech-to-Text, Translate, and Text-to-Speech APIs: https://daleonai.com/translate-dub-videos-with-ml
You can read the full guide on the Google Cloud blog below:
Your Ultimate Guide To Speech On Google Cloud
New Google API Helps You Win The Last Mile
If you’re like most people, when you go online and order a pair of house slippers shaped like adorable cartoon lobsters, you probably don’t give much thought to how the slippers make their way from the cartoon lobster slipper factory to your doorstep.
But getting you your lobster slippers on time, along with millions of other customers’ deliveries shipped each day, is a miraculous feat of modern logistics engineering.
And for businesses, the stakes couldn’t be higher. In our always-on, on-demand world, customers expect to get whatever they want, whenever they want it — like that kid in that episode of The Twilight Zone that got spoofed in that episode of The Simpsons. You know, the one where Bart turns Homer into a Jack-in-the-box.
Thankfully, Google Cloud recently announced the public general availability of the Cloud Fleet Routing(CFR) API.
With CFR, businesses can harness the power and scale of Google Cloud’s infrastructure, solving route planning problems in seconds, and helping to transform the Bad Old Days of last-mile logistics…
(Life before the Cloud Fleet Routing API)
…Into the Good New Days of last-mile logistics.
As Christopher Cho, Product Manager, Cloud AI and Industry Solutions, writes in the article announcing the public release of the CFR API:
Last mile operations are fraught with disruptions: change of delivery time, vehicle malfunction, last-minute workforce changes, and the list goes on. To help our customers stay agile, the CFR API lets them reoptimize their existing plan up to 20 times a day without incurring any additional cost. This ability is especially useful for the last mile, where dynamic changes such as traffic congestion or ad hoc requests can require frequent re-optimization. As a result, customers can use our CFR API to plan effective disruption responses worry-free, so that operators can quickly take the best action with consumers as soon as possible.
Because, let’s face it: Nobody should be without their cartoon lobster slippers. And now, with the Cloud Fleet Routing API, no one will ever have to be.
Get the full post here:
Moloco Scales Cloud Bigtable to 5M+ Ad Requests Per Second
Moloco is an AI company that builds custom machine learning models to help customers predict and optimize the performance of their mobile ads.
This isn’t as easy as it sounds. [Editor’s note: It doesn’t sound particularly easy]
For one thing, most ad exchanges put 150 millisecond time limits on bid requests, so Moloco loses its bid opportunity if it misses that window.
As Moloco VP of Engineering Chang Kim says on the Google Cloud Blog, building out an in-house technical infrastructure that would allow Moloco to “process the request, sort campaigns, apply ML models for inference to evaluate campaigns against the request, adjust our predictions based on inventory types and bid floors, and finally return bid responses”, would have been prohibitively time-consuming.
Instead, Moloco chose to build their ad tech platform using Google Cloud services including Bigtable, Compute Engine, Cloud Storage, Google Kubernetes Engine, and Vertex AI. This lets Moloco build fast data pipelines and get insights in real-time.
By building our ad tech platform using Google Cloud, we’ve expanded the number of bid requests we can handle, from 550,000 to over 5 million per second, that’s an increase of more than 9 times from where we started about two years ago.
It was unclear as of this writing if Moloco has plans to develop a computerized, GPT-4 Enabled Don Draper simulation who lives in the Cloud to actually write the ad campaigns for clients, but I wouldn’t be surprised if it’s in the works.
Moloco Handles 5 Million+ Ad Requests Per Second With Cloud Bigtable
Keeping Data Secure With Security-By-Design
It seems like not a week goes by without another headline-grabbing security breach by cybercriminals. The past few years alone have seen the Equifax data breach, Crypto exchanges compromised to to tune of billions of dollars, and the massive SolarWinds hack (which affected thousands of public and private organizations including Microsoft, Cisco, and the US Department of Homeland Security).
With no end in sight, and with the consequences of these attacks ranging from identity theft and fraud, to national security concerns, it’s more important than ever for organizations to stay ahead of attackers.
The bad news is that hackers don’t sleep (it’s all that Red Bull), and keeping ahead of them requires speed and resources that aren’t always available or made a top priority.
In this article, M.K. Palmore of Google Cloud’s Office of the CISO, and Wade Holmes of Google Cloud’s security division, describe the problem:
A close look at the root problems behind high-profile security breaches reveals that it’s a lack of agility and an inability to scale resources that prohibit the modern security organization’s ability to respond quickly enough to counter new challenges. Look even closer and you’ll often find an insufficient implementation of best practices and ineffective solutions, leaving an organization continually chasing the next tool or solution and scrambling to stay ahead of emerging threats.
The solution starts with building security features into cloud infrastructure from the ground up, including best practices like encryption and pre-defined user roles with limited rights as the default.
Strong security in the public cloud starts with the foundational pieces: the hardware and design elements. At Google, for example, we take a security-by-design approach within both the data center and purpose-built components themselves. Within Google Cloud, data is encrypted by default — both at rest and in transit. Google’s baseline security architecture adheres to the zero trust principles, meaning that every network, device, person, or service initially cannot be trusted.
Some of the top benefits of Google’s security-by-design approach are reduced risk, reduced cost, and helping companies stay off of Buzzfeed’s Which Famous Preventable 2020's Corporate Data Breach Are You? Quiz.
(I’m an Equifax, if you were wondering)
Anyway, that’s all for this week’s edition of Cloudnomics. But stay vigilant out there.
Because I can guarantee you that somewhere, right now, there’s a guy sitting at a computer screen, with his sights set on your company’s most sensitive data.
Don’t let this happen to you:
(“Oh. My. God. They forgot to configure multifactor authentication!”)
Full article below:
The journey to the cloud mitigates enterprise risk
P.S. If you enjoyed reading this, follow Cloudnomics for more on how real-world organizations are using The Google Cloud Platform and Machine Learning tools to achieve impressive results.
Top comments (0)