Google DeepMind shifts from research lab to AI product factory
Google DeepMind shifts from research lab to AI product factory
Why Big Tech is fine with launching unreliable AI apps
Google DeepMind shifts from research lab to AI product factory (link)
Why Big Tech is fine with launching unreliable AI apps (link)
Microsoft updates Recall feature after security and privacy backlash (link)
Grow your business with viral clips from Munch - at 50% OFF
Staying ahead in the social media landscape is a challenge that every business owner, marketer, and agency faces. Munch offers a groundbreaking solution, leveraging AI to transform your long-form videos into engaging, bite-sized clips primed for social media sharing.
Limited time offer: 50% OFF for the 1st month on all Munch plans - use code SMB50 at checkout to save!
***
Google DeepMind shifts from research lab to AI product factory
In a span of just one week in mid-May, two tech giants made significant announcements about artificial intelligence products built on a groundbreaking technology developed by Google. On May 13, OpenAI Inc. unveiled an updated version of the model powering ChatGPT, its incredibly popular chatbot, which relies on a technology called a transformer that Google first introduced in a 2017 research paper. The very next day, Google launched AI Overviews, a feature that provides search results with answers generated by its own system, also based on the same transformer technology.
However, the AI Overviews launch was marred by embarrassing mistakes, such as suggesting that people eat rocks and put glue on pizza. In response, Google implemented new safeguards just a week later, explaining that they were designed to prevent the feature from presenting satirical content as factual information. This misstep was particularly damaging for Google, which is under immense pressure to keep up with competitors like OpenAI. Many in the tech industry believe that products like ChatGPT have the potential to render Google Search obsolete, which would be a devastating blow to the company's revenue stream.
In an effort to stay ahead of the curve, Google announced in April 2023 that it would merge its two elite AI teams, Google Brain and DeepMind, into a single "super-unit" called Google DeepMind. This new unit faces the daunting task of improving Google's commercial AI products while maintaining the company's historic strength in foundational research.
The catalyst for this shift was OpenAI's 2020 paper on its GPT-3 model, which prompted Google to reassess its approach to AI development. If a startup with fewer resources could leverage Google's own research to gain a competitive edge, perhaps the company needed to rethink its strategy….
Originally by Julia Love and Mark Bergen of Bloomberg Businessweek
***
Why Big Tech is fine with launching unreliable AI apps
The rocky debut of Google's AI-generated search summaries in mid-May was, in many ways, a predictable outcome. And it's likely to happen again.
A pattern has emerged in the high-profile launches of generative artificial intelligence applications from tech giants. Initially, these products generate significant buzz, but once they're in the hands of users, the excitement quickly gives way to disappointment as errors and unreliability come to light. This is often followed by a wave of negative media coverage, occasionally accompanied by an explanation or apology from the company.
It appears that some companies are willing to release flashy but incomplete generative AI products under intense competitive pressure, even if it means risking embarrassment. The concept of a minimum viable product, where a company releases a basic application to gauge customer needs and demand before developing a full-featured version, has been a staple in the tech industry for years. However, generative AI companies are pushing this concept to the limit…,
Originally by Joe Castaldo of The Globe and Mail
***
Microsoft updates Recall feature after security and privacy backlash
Microsoft has announced a series of changes to its controversial "Recall" feature, following criticism from security and privacy experts. The AI-powered tool, which was unveiled last month as part of the company's Copilot+ PCs initiative, has been revamped to become an opt-in experience, with additional security layers added to address concerns.
Recall, which takes regular screenshots of user activity on a machine and creates an index that can be queried using AI, sparked backlash from security and privacy leaders. In response, Microsoft is updating the set-up experience of Copilot+ PCs to give users a clearer choice to opt-in to saving snapshots using Recall. If users don't actively choose to turn it on, the feature will be off by default.
In a blog post on Friday, Microsoft executive Pavan Davuluri outlined the changes, which include the addition of new safeguards. These include proof of presence to view a user's timeline and search in Recall, as well as "just in time" decryption. The Copilot+ PCs initiative combines Microsoft's Copilot artificial intelligence technologies with its Windows PC operating system….