Google goes full force with AI announcements, IBM comes out swinging and disses other AI offering as inaccurate and presenting its offerings as more responsible, and could we see the death of the standard hard drive?
 |
 |
 |
These top tech news stories and more for Thursday, May 11th, 2023, I’m your host Jim Love, CIO of IT World Canada, and Tech News Day in the US.
There were a raft of new developments unveiled at Google’s I/O developer conference on Wednesday.
Google’s CEO Sundar Pichai announced Palm2, a more sophisticated Large Language Model it will be using in more than 25 AI powered products.
Google says Palm2 will come in a variety of models and sizes, one of which can run natively on a mobile device. Google is also using the new model in search, Google Cloud and even its Bard chatbot.
Clearly Google was hoping that the conference would allow them to show that they were still competitive in the AI race, and hopefully a smooth “relaunch” would allow them to bury the disappointing Bard launch early this year.
The new and improved Bard will supposedly have improved math and reasoning capabilities and be more helpful in writing code.
It is also slated to have images and videos in its results and even to allow images in its prompts.
Google is abolishing the waitlist and offering Bard in English in 180 countries. It is expected to rapidly scale the number of languages supported to 40, starting with Japanese and Korean.
It claims that it will add generative AI into its search results, as well, and showed demos that integrate video, pictures and weblinks in a chat style format.
Google also announced plans to incorporate ads into its generative AI results. How they are going to do that is another question and Google says it is experimenting with how to serve ads with generative AI results. In the meantime, it is making AI search results available in its experimental labs program.
And they didn’t stop there. Everything at Google got some kind of AI boost. Gmail and Google Workspace will be enhanced with AI driven features. Pichai demonstrated a “Help me write” feature with an example of how, for instance, you write an email to seek a refund from an airline, automatically finding flight details from prior emails.
Google is also adding AI to its spreadsheets. These will be part of a new service called Duet which will be tested and rolled out later this year.
Google Cloud will be enhanced with AI with it’s Vertex offering which allows users to, in Google’s words, “Build, deploy, and scale machine learning (ML) models faster, with fully managed ML tools for any use case.” Google is apparently adding models that can be used for coding image generation and speech recognition.
And oh, yeah, they have some new Pixel devices, including a foldable smartphone and a long-awaited Pixel tablet. Ever think you would see the time when a new phone design was buried at the bottom of a list?
As Pichai said, “As you might have heard, AI is having a big year.”
Sources include: Axios
IBM has come out swinging, claiming its latest AI offering, which it will bring to market in July, is far more accurate and more responsible that their rivals.
And where Microsoft, OpenAI and Google are trying to win large consumer markets with their Generative AI Solutions, IBM is focusing on companies and offering AI to IBM clients with its “data model factory” and specialty products tuned to specific industry needs.
In fact, Dario Gil, IBM’s head of research, quoted in Axios as speaking at the Watsonx preview on Monday was only one of the IBM execs to diss their competition, saying that system like ChatGPT are “not ready for primetime” because they have “all sorts of random and made-up facts.”
Is this an accurate competitive advantage? Or is it just IBMs way of justifying that it has lost the lead in AI, at least in the public consciousness.
IBM was once the absolute leader in AI and everyone remembers how its original AI system, named Watson, won Jeopardy in 2011. But the company didn’t find huge success, in fact, Axios reported that its revenue declined for 10 consecutive years in this area.
Today, IBM is struggling for relevancy against their hugely popular competitors. Can it find its way back to leadership by throwing shade at the competition?
CEO Arvind Krishna wasn’t at last week’s AI meeting at the White House. Krishna told Axios, he was glad because it gave him more time to court clients“ who care a lot about accuracy.”
Yeah. Who wants to be invited to the White House to meet the President and advise him on the future of a key technology.
Sources include: Axios
Microsoft Azure’s CTO Mark Russinovich is on a mission to promote “confidential computing.”
Confidential computing happens in a “trusted execution environment” using hardware-based security mechanisms to protect any code from everything outside its secure environment.
At least the first generation of confidential computing is based on Software Guard Extensions (SGX) technology built into Intel’s Xeon processors.
Early models required code to be reworked to run in the confidential “enclave” but new technology allows users to “lift and shift” workloads without alteration in to a protected virtual machine.
What’s driving this new technology? One thing is that it allows a number of users to bring all their data together in “data clean rooms” where it can be analyzed and processed securely and give companies a lot more that they could get just analyzing their own data.
The new tech removes some of the obstacles that would prevent users and companies from sharing sensitive data.
According to Microsoft, the Royal Bank of Canada (RBC) has already established a clean room where they combine their data and purchasing data from merchants and get a much fuller view of purchasing behaviour without seeing or exposing any confidential information that consumers might share with the merchants.
Similar models are being envisioned for sharing of health and medical information.
It’s generating a lot of interest from regulated industries in the public sector, healthcare and financial services.
Russinovich claims that in the future, confidential computing will no longer be a specialty feature, but it will be the default way of processing information.
Sources include: The Register
According to Shawn Rosemarin, VP of R&D in the Customer Engineering unit at Pure, no more hard disk drives will be sold after 2028 being replaced by solid state drives and flash storage.
Why? According to Rosemarin, power is one answer. He has said, “Our CEO in many recent events has quoted that 3 percent of the world’s power is in datacenters. Roughly a third of that is storage. Almost all of that is spinning disk. So if I can eliminate the spinning disk, and I can move to flash, and I can in essence reduce the power consumption by 80 or 90 percent while moving density by orders of magnitude in an environment where NAND pricing continues to fall, it’s all becoming evident that hard drives go away.”
He points out that AI will also play a huge role. All our projections about data growth were done before the explosive growth of AI, he says – and with the number of video and audio files that will be created by AI next year, the situation gets worse.
Suffice it to say that note everyone agrees with him. But hard drives are still a very popular data storage medium, relatively cheap and always getting faster. Gartner claims that enterprise SSDs will reach 35 per cent of storage by 2026, making Rosemarin’s 2028 date, unlikely.
Sources include: Blocks and Files
And a suburb of Montreal, Canada is trying an experiment with traffic lights that will turn green for the good drivers who keep to the speed limit.
The lights are designed to stay red until it senses an oncoming car and only change to green if the car is going the speed limit.
These types of lights are used in Europe but have never been tried before in North America.
And if you’ve ever driven in Montreal, you may be asking, tongue in cheek, who the two drivers were that actually slowed down..
But the light does more than slow down drivers. It also reports statistical information about drivers and their behaviour.
And that data shows some impressive results. In the test, albeit on a suburban and quiet street, the average vehicle speed dropped by more than 25 per cent and whether you express that in kilometres or miles per hour – that’s an impressive result.
Can technology kill the speedbump? Here’s hoping.
Sources include: Jalopnik
That’s the top tech news for today. We go to air with a daily newscast five days a week, as well as a special weekend interview with an expert on topics relevant to today’s tech news.
Follow Hashtag Trending on Google, Apple, Spotify or wherever you get your podcasts. And you can even get us on your Alexa or Google smart speaker. You can even find us on YouTube as TechNewsDay.
We love your comments. You can find me on LinkedIn, Twitter, or on Mastodon as @therealjimlove on our Mastodon site technews.social. Or if that’s too much, just leave a comment under the text version at itworldcanada.com/podcasts and you can find all of the links in those text versions.
I’m your host, Jim Love. Have a Thrilling Thursday!