Debating the Responsibility for Creating AI Tools and Analyzing Sales Funnel Data

This week Brian sits down with our own AI expert, Jordan Ranous, to discuss and speculate on the challenges facing enterprises and small businesses as they try to get a grip on the hot topic of the decade, Artificial Intelligence. If you have been following StorageReview, then you are aware of some of the tests, reviews, and trials we have been exploring in our lab. Jordan is engaged in almost all AI content posted on our site.

This week Brian sits down with our own AI expert, Jordan Ranous, to discuss and speculate on the challenges facing enterprises and small businesses as they try to get a grip on the hot topic of the decade, Artificial Intelligence. If you have been following StorageReview, then you are aware of some of the tests, reviews, and trials we have been exploring in our lab. Jordan is engaged in almost all AI content posted on our site.

Jordan is an experienced AI technologist with a background working with financial institutions. Brian always offers his guests some tough questions, and Jordan gets the same treatment. AI is such a hot topic, and the buzz in the industry just keeps getting louder, generating more questions than answers. Jordan has a rare talent of being an expert in AI and being able to articulate answers that can be easily understood.

This is the first podcast that we’re actually integrating live with our Discord community so they can interact and ask questions on topics we are hitting at that moment.

If it’s AI-related, Brian and Jordan talk about it. It is certainly worth watching the podcast in its entirety, but if you would like to hop around, we have the time-stamped transcript below.

Join the Discord to participate in future pods in real-time.

00:00 Introduction

• Setting the stage for the live event
  • Invited the StorageReveiw Discord community
  • Exploring AI in the lab
  • Hands-on
• Recapping Flash Memory Summit (FMS
  • AI Vision Demo
  • SR Sloths stole the show
• AI does not have a universal definition
  • Generative AI tools
    • ChatGPT
    • Dall-E
• Generative AI vs AGI (artificial general intelligence)
• Computer Vision
  • A subset of broader AI
  • Object recognition
  • Image Datasets

05:00  Finetuning the model

• Business Intelligence
  • Categorizing images
  • Identifying brands rather than just shapes
  • Self-checkout
  • Loss Prevention
• Creating the Datasets
  • Training data
  • Install disparity between enterprise and small retail shops
• Tools for AI
  • Leveraging the right tools for the job
• Implementation challenges
• Creating chatbots for various industries

10:00 Chatbots for customer service

• Chatbot is programmed to fix the problem, not initiate transferring the call
• Pressing zero doesn’t do anything
• Creating higher frustration for customers
• Catching AI doing something it shouldn’t
• AI has the potential to hyperextend in any direction
  • Keep an eye on your AI model to keep it honest and focused
  • Quality data means quality results

15:00 Security is paramount

• Inherent biases in AI
• Training data collection is critical
• Gather a diverse group to build the training data
• Proper finetuning
• Be careful of the persona you create
• Doom server demo
  • AI’s talking trash
  • AI gamers fighting each other and making disparaging jokes
  • AI understood the assignment and acted upon it
• Ethics
• Enterprises might be reluctant to provide too much information
  • Know where it should go
  • And where it might go
  • The fear of the unknown

20:00 Active Directory for access levels

• Keeping data secure and available to appropriate teams
  • Call centers probably do not need access to proprietary marketing material
  • NeMo Guardrails
  • How questions are asked or how to prompt the question
• Hiring Prompt Engineers
• Support tools
  • Democratizing the toolset

25:00 What’s the on-ramp for small businesses

• Set up internally
• Create solid prompts
• Do the simple stuff first
• Ease into the technology
  • Understand how it works
  • Understand limitations
• Use an AI model to analyze CRM data
  • Help with sales forecasts
  • Analyzing the sales funnel
• Use built-in tools
• Not every business needs a custom model

30:00 What’s a good starting point

• Fund IT for AI APIs
• Consume available tools
• Employ a paid version of ChatGPT4
• Monitor billing
• API provides additional freedom

35:00 What about the hardware

• Workstations with GPUs
• Keep GPUs busy
• Parallelize
• Rack space consideration
• Liquid or Air cooled
• TBs of storage
• GPUs take slots that might have been used for storage
• How much power can a rack support

40:00 Dell’s latest CPU servers

• Fast, shared storage
• Software
•  SSDs
  • E1.S
• Inferencing cards
• Ultra-high-speed servers and workstations

45:00 Where cloud can benefit AI

• OVH
• Cloud is a valuable resource
• Hourly billing

50:00 Wrap-up

<span data-mce-type="bookmark" style="display: inline-block; width: 0px; overflow: hidden; line-height: 0;" class="mce_SELRES_start"></span> Podcast video

StorageReview Discord

Dell XE9680 Review

Subscribe to our podcast:

• iTunes
• Spotify
• Google Play

Engage with StorageReview

Newsletter | YouTube | Podcast iTunes/Spotify | Instagram | Twitter | TikTok | RSS Feed

The post Podcast #122: Navigating The AI Landscape: Real-World Insights And Challenges appeared first on StorageReview.com.