Accelerate your knowledge now

AI IS ACCELERATING - GET MOVING

Let's help you understand the WTF of AI and how it helps YOU and your organisation grow. Ready, set, fall.

This week:

?? AI Goes Wrong: Apple Intelligence Is Brutal

?? Elite Resources: What AI Is And Isn't

?? AI Made Simple: Guardrails Vs Wild West

?? Tips & Tricks: Google Learn About

AI GOING WRONG: Messages Can Be Brutal

As part of the Apple Intelligence rollout (and iOS 18) AI notification summaries are here - and they've got a mean streak.

Whilst notifications can be handy, these AI ones come unstuck when trying to summarise group chats, complex subjects or being dumped while you were snoring.

Apple has squeezed in a mini-LLM onto its devices, which tries its best to grasp the nuances of human communication and often fails miserably. Here are some more Apple 'Intelligence' notification gems.

"Apology for subpar communication, not ready for relationship" Has to be a dev, right?

"Multiple people at front, back and side door" Drug raid on your gaff? Forgotten you're party host tonight?

"Tupac turns up at the end, insane Bitcoin rise" Redemption arc? Death Row manipulating Crypto?

Get in touch to talk about AI and your organisation, we promise not to be subpar or dump you while you're in the the land of nod.

AI RESOURCES: What AI Is And Isn't

Generative AI is often oversimplified as a straightforward process: feed data in, and instant outputs follow. But this perception barely scratches the surface of what AI entails and the challenges it brings:

What people think AI is:

• Feed in some data.
• The AI figures it out.
• Get instant insights, analysis, or content.

What AI actually looks like:

• Data: A detailed process of prep, analysis, cleaning and engineering.
• Model Development: A mix of selection, training, and evaluation.
• Refinement: An iterative process that improves over time.

There are also some noteworthy constraints. Legal, ethical/transparency, historical bias, and security.

• Could AI be used to deceive with deepfakes?
• Could AI be used to imitate the voice of a loved one?
• Could AI be using damaging stereotypes?

Use the sheet to check your understanding and move forward with more knowledge, you can also get free AI training here.

AI MADE SIMPLE: AI Guardrails Vs Openness

As the AI landscape evolves, so does the debate over how much freedom large language models (LLMs) should have in outputting content.

With models like Musk’s Grok and the newly launched Hermes 3, the choices are wider than ever for businesses.

Here’s the rundown on some popular low-moderation LLMs:

• Minimal Guardrails with Grok: Elon Musk’s LLM, Grok, has limited moderation and a neat line in sarcasm, pushing the boundaries of what AI assistants can say and create.
• Balanced Moderation by Hermes 3: Nous Research's Hermes 3 fine-tunes Meta’s Llama 3.1 for a (much) more relaxed moderation policy - providing a middle ground between rigid and open AI, offering users more flexibility.
• Mixed Models Like Mixtral: Some platforms, like Mixtral, opt for adaptive moderation based on user context, balancing openness with safety.

The Potential Risks: Lower moderation can allow for more creative and engaging outputs but increases the risk of misuse or harmful content generation.

What does it mean for businesses? Open models like Grok or Hermes may suit companies aiming for edgy marketing or bold communication styles but require careful oversight. On the other hand, more moderated models ensure safer interactions but could limit creativity.

Looking to choose the right AI model for your business? Partner with G3NR8 to explore the most suitable LLM options tailored to your company’s needs.

TIPS & TRICKS: Google Learn About

Google is combining search and education with 'Learn About', its latest AI experiment - offering a more educational experience with detailed, visual responses rather than basic snippets.

Key Points:

• More than answers: Learn About delivers structured explanations with visuals and defined terms, aiming for deeper insight rather than top-level (i.e. blue link) responses.
• Focus areas: Currently tailored to topics like science and history, suggesting an initial push toward aiding students and curious learners.
• User experience shift: Encourages users to stay within Google’s ecosystem, potentially impacting traffic to educational sites.
• Educational Model: it's based on Google's LearnML model that's grounded in educational research and learning patterns.

What it means: This move could redefine how people engage with learning online, giving Google more influence over educational narratives.

While it may open up access to learning, it centralises it within Google's eco-system and that will raise concerns over bias and content diversity.

WHATS NEW - Trends and Insight Shorts

> AlphaFold goes Open Source

> OpenAI win latest copyright round

> Get free Washington Post content

> ￡1m for Turing AI Art

> AI wins its first Grammy

G3NR8 is an impact consultancy for the exponential age. We help organisations figure out the highest impact areas to use AI for maximum impact.

Work with us to:

• Develop your AI strategy and policies.
• Create a competitive advantage using your proprietary data.
• Automate repetitive tasks.
• Gain actionable insight from data.
• Train your teams - we've trained 120+ large organisations in advanced generative principles and implementation.

Drop us a line for a chat about productivity gains, cost savings and how AI can power growth in your business.