Can’t focus? Maybe you have TikTok brain. It’s official – AI has reached the top of the hype curve according to Gartner. LinkedIn is providing a way to prove that we are who we say we are and if walks like a Platypus and talks like a Platypus, it may be a new open source AI model and one that is gaining attention for more than its unusual name.
 |
 |
 |
These stories and more as we bring you the top tech news stories on today’s Hashtag Trending.
I’m your host Jim Love, CIO of IT World Canada and Tech News Day in the US.
In a recent study, researchers explored the influence of TikTok on attention spans, particularly among children. The study encompassed participants across different age groups, examining their app usage habits. Notably, the study revealed that prolonged engagement with TikTok was linked to reduced attention spans, especially in children.
This presents a potential hazard as diminished attention spans could impact children’s ability to focus on tasks like schoolwork and even interpersonal interactions.
The study referenced an interesting point from the original article about brain development. It was highlighted that the prefrontal cortex, responsible for cognitive functions such as impulse control and decision-making, isn’t fully developed until around age 25. This underscores the vulnerability of younger users to the potential consequences of extended digital engagement.
Sources include: The Week
Generative AI has claimed a spot on Gartner’s influential Hype Cycle for Emerging Technologies for 2023, according to a recent announcement from the firm. This subset of AI is positioned on the “Peak of Inflated Expectations” within the cycle, signaling significant industry buzz and high expectations.
Melissa Davis, a Gartner vice president analyst, explained that the Peak of Inflated Expectations marks a phase where excitement surpasses the technology’s current capabilities.
Despite this, Gartner predicts that generative AI will deliver “transformational benefits” in the coming two to five years.
Sources include: Tech Republic
I don’t know if you remember the huge wave of fake identities that were appearing on LinkedIn a few years back? It slowed down, but I still see them – people who want to link up with you but are somehow suspicious?
Starting August 17th, a new avenue for identity verification arrives for Canadians on LinkedIn. It’s already in the US. By integrating CLEAR Verified, a state-of-the-art verification platform, LinkedIn members can now bolster their profiles with an additional layer of authenticity. Utilizing their Canadian government ID and their phone number, members can validate their identity through CLEAR. LinkedIn claims this will empower Canadian members to forge a network founded on trust, and enhance the safety and dependability of the platform for all users. For how to set up the process, consult the LinkedIn Help Center.
Sources include: LinkedIn
A cybersecurity advisory panel in the US is set to investigate the targeting of cloud computing environments, with a specific focus on Microsoft’s involvement in a recent breach of government officials’ email accounts by suspected Chinese hackers.
The Cyber Safety Review Board, established by the Biden administration, will examine strategies that cloud service providers, the government, and the industry can adopt to enhance identity management and authentication in cloud environments. This initiative comes in response to the increasing reliance on cloud computing for critical services. The decision to scrutinize cloud security and Microsoft’s role in the breach follows a request by Senator Ron Wyden, who seeks accountability for cybersecurity practices. Microsoft’s transparency and security measures are apparently under heightened scrutiny.
Source: Data Centre Knowledge
As the demand for digital services surges, data centers are grappling with their substantial energy consumption. The International Energy Agency estimates that in 2022, global data centers used 240 to 340 terawatt-hours (TWh) of electricity, equating to 1 per cent to 1.3 per cent of global electricity demand. This excludes the energy consumed by cryptocurrency mining. To combat emissions and align with the Net Zero Emissions scenario, a 50 per cent reduction by 2030 is vital.
Governments are looking at ways to regulate energy use in this growing sector.
Germany is considering the Energy Efficiency Act, compelling public authorities, corporations, and data centers to reduce energy usage for sustained climate protection and lower energy consumption in Germany.
The threat of regulation is prompting a search for creative solutions. In addition to working on hardware that consumes less energy both in processing and storage, some data centres are trying new approaches like reusing what would be waste heat to provide heating to nearby buildings.
Source: Data Center Knowledge
In the realm of artificial intelligence, a new and unusually named model is gaining attention. Platypus 2 70B AI has emerged as a leading open-source large language model (LLM), securing the top position on HuggingFace’s Open Large Language Model Leaderboard.
Central to Platypus’s success is the introduction of the Open-Platypus dataset. Carefully curated from various open datasets, this publicly accessible dataset played a pivotal role in fine-tuning and merging what are called Long Range Arena or LoRA modules which extend the amount of information that a model can consider and handle . This unique approach enables the model to retain the strengths of pre-trained LLMs while allowing for fine tuning and more specific domain expertise.
The Platypus models, built upon the LLaMA and LLaMa-2 transformer architectures, leverage LoRA and PEFT Parameter Efficient Fine Tuning for enhanced performance.
Platypus’s excellence is demonstrated by its impressive quantitative LLM metrics across various model sizes. Remarkably, it achieves this efficiency with fewer fine-tuning data and compute resources compared to other state-of-the-art LLMs. Notably, a 13B Platypus model can be trained on a single A100 GPU using 25 thousand questions in just 5 hours.
The research’s focus on optimizing LLMs through PEFT and LoRA, combined with the STEM and logic-oriented Open-Platypus dataset, presents a remarkable leap in AI advancements. The approach’s efficiency, paired with meticulous data handling, propels Platypus 2 AI LLM to the forefront of AI research.
Source: Geeky Gadgets
Those are the top tech news stories for today. Hashtag Trending goes to air 5 days a week with a special weekend interview show we call “the Weekend Edition.”
You can get us anywhere you get audio podcasts and there is a copy of the show notes at itworldcanada.com/podcasts
If you want to catch up on these and other news more quickly, you can read these stories and more at TechNewsDay.com and ITWorldCanada.com on the home page.
It’s always great to hear from you. It lets me know we are on the right track. Or possibly, on the wrong track. Or are we a mixture of two things – sort of a platypus of journalism?
So please go to the article at itworldcanada.com/podcasts – and you’ll find a text edition there. Click on the x or the check mark but tell us what you think. You don’t even have to show ID – we’ll trust you.
To those who have reached out – my sincere thanks.
I answer each and every email that comes to me. It is so great to hear from you.
I’m your host, Jim Love. Have a Fabulous Friday!