Everyday AI Podcast – EP 423: AI News That Matters - December 16th, 2024

Main Theme & Purpose

In this energetic roundup, host Jordan Wilson unpacks the week’s most significant AI stories, focusing especially on the rapid-fire rivalry between OpenAI and Google, while also touching on key developments from Apple, Klarna, and the broader AI landscape. The episode is structured as a fast-paced briefing to keep listeners ahead of the ever-accelerating AI curve and make practical connections to business and career growth.

Episode Structure

• OpenAI news avalanche (including new tools, features, and integrations)
• Other notable AI stories: Apple, Klarna, Devin, Microsoft’s Phi-4, and more
• Google’s massive announcement spree (Gemini 2.0, Project Astra, new agents, and more)

Key Discussion Points & Insights

OpenAI News Highlights (02:14 – 17:00)

1. OpenAI Sora: Text-to-Video Generator (03:09)

   • Sora, OpenAI’s new tool, creates videos from text prompts.
   • Available for ChatGPT Plus ($20) & Pro ($200) users: Only Pro users can generate human videos.
   • Notable: Signups reopened after prior server overwhelm.
   • “Sora is wiping the competition... by far the best ones are also going to be Sora. But this is the worst it will ever be.” (04:43, Jordan Wilson)
   • Compared to established rivals (Runway, Kling, Luma Labs), Sora is noted for its unprecedented generative ceiling.

2. ChatGPT Canvas Expanded (06:10)

   • Canvas, a real-time collaborative workspace for content editing, is now available to all users, including free tier.
   • Adds ability to run Python directly in chat (previously only cloud artifacts allowed in-depth browser coding).
   • “Go ahead and upload a spreadsheet in there... say like hey, create me a visual that could be helpful. Create a dashboard that could be helpful and it will do it.” (10:00)

3. Advanced Voice Mode Gains Video & Screen Share (11:20)

   • Now includes vision: ChatGPT can “see” via phone camera—think: virtual whiteboards, math homework, live business brainstorming.
   • User Remarks: “I love the video feature but haven’t discovered much usefulness to it yet… so far just used it for fun.” (13:05, Michael)
   • “It remembered my dog’s names.” (13:15, Dr. Harvey Castro)

4. ChatGPT ‘Projects’ Feature—Folders & Metadata (14:00)

   • Users can organize files and chats via folders, reminiscent of Anthropic Claude’s projects and Google’s Notebook LM.
   • Early testing shows better document retrieval than custom GPTs for some tasks.

5. Apple Intelligence Officially Taps ChatGPT (17:00)

   • Apple’s “significant update” integrates ChatGPT into Siri for more complex queries, now active on iOS 18.2 and Mac 15.2.
   • Jordan: “It’s a nothing burger... all stuff that we had through other tools pre-ChatGPT…” (18:57)

AI News Beyond OpenAI & Google (19:50 – 32:55)

1. Cheaper iPhone for Apple Intelligence (20:40)

   • Rumored iPhone SE4 to feature Apple’s A18 chip, enabling advanced AI features at a lower price (potentially $499–$599).

2. Eric Schmidt (Former Google CEO) Warns of Runaway AI (22:18)

   • Cautions on self-improving AIs; proposes dual-system approach—one monitoring the other for safety.
   • “AI systems capable of independent decision making could emerge within two to four years.” (23:50, paraphrased)

3. Klarna Embraces AI Over Human Rehiring (25:15)

   • Klarna’s CEO Sebastian Siemiatkowski touts workforce reduction from 4,500 to 3,500—attrition is now mostly filled by AI, not new hires.
   • “We’re just not hiring people anymore... as many human roles over to AI as possible.” (26:00)

4. Devin by Cognition Releases, $500/month (28:26)

   • Positioned as a junior developer AI; handles ongoing engineering tasks without frequent human prompting.
   • Integrates with Slack, IDEs, APIs; intended for code cleanup, bug fixing, drafting press releases, etc.

5. Microsoft Phi-4, Mini Model—Big Performance (30:49)

   • New 14B parameter model reportedly outperforms much-larger rivals in math and reasoning.
   • “Small models... the future is hundreds of specialized small models.” (31:30)

Google's AI Avalanche (33:04 – 48:05)

1. Gemini 2.0 Launches; Agentic Era Beckons (33:30)

   • Debuts with “Flash” model (cheap, fast), already besting old larger models in benchmarks.
   • “Even though it is a flash model... this thing is a banger.” (34:50)

2. Google Agent Space for Enterprise AI Agents (36:12)

   • New cloud platform for branded company AI agents, integrates with Google/Microsoft tools.
   • “Google goes typical go-to-market... you can sign up for early access—but will you get it? Probably not.” (37:30)

3. Project Astra Progress—AI Glasses and Physical World Assistants (39:17)

   • Demonstrated but still unavailable.
   • Hint at future where persistent on-body agents (glasses, XR) assist contextually.
   • User Opinions: Divided on return of AR glasses—“I don’t know how realistic [it is].” (41:17)

4. Android XR: Extended Reality Collaboration with Samsung, Qualcomm (43:07)

   • Platform bringing Gemini to “glasses and headsets.”
   • Early launch with Samsung Project Muhan expected next year, aims at much cheaper/lighter alternative to Apple Vision Pro.

5. Google Notebook LM—Paid Plan & Audio ‘Call-In’ (44:32)

   • “Call-In” feature lets users interrupt AI podcast-mode overviews for real-time Q&A.
   • “This one little feature... I think is one of the most exciting small features that everyone is overlooking.” (46:09)
   • Paid (Plus) tier enables team sharing, higher limits, and new content creation pane.

6. Project Mariner: Autonomous Chrome Agent (47:20)

   • Chrome extension lets Gemini perform tasks for you online (e.g. research, shopping).
   • Competes with Anthropic’s “computer use,” but more accessible.

7. Deep Research: Google’s Perplexity Rival (48:20)

   • Tool within Gemini Advanced for generating in-depth, multi-source reports.
   • Jordan: “It visited 169 websites... took about 2 or 3 minutes. Can you imagine that?” (49:01)
   • Direct shot at Perplexity’s crown for AI research assistants.

8. Google AI Studio: Real-Time Multimodal Voice & Vision (49:44)

   • Users can now interact with Gemini via live voice/vision for instant answers and analysis, both desktop and mobile.
   • No waitlist; first fully-integrated, vision-enabled voice mode among major AI providers.

Notable Quotes & Memorable Moments

• On Sora’s capabilities: “Sora is wiping the competition... Sora by far has the highest, highest ceiling... This is the worst it will ever be.” (04:43, Jordan Wilson)
• On Apple’s ChatGPT/Siri integration: “Siri is the middleman or middlewoman now. Just, you know, passing off all our queries to... ChatGPT. That’s funny.” (19:45, Jordan referencing audience comment)
• On AI workforce disruption: “Klarna was one of the big, you could say AI case studies of just essentially unabashedly handing over human roles, being like, nope, nope, we’re not hiring humans anymore.” (26:44)
• On Google and AI agents: “Google did not really create any marketing, any messaging, seemingly any real strategy... and then the middle of last week, Google went bananas. B A N A N A S. Like they went wild.” (33:08)
• On Google’s Notebook LM ‘call in’: “This one little feature... I think is one of the most exciting small features that everyone is overlooking. So don’t sleep on that.” (46:09)
• On Gemini vs. Perplexity in research: “I use this to help me research one of my shows last week. A single prompt inside... it visited 169 websites... took about 2 or 3 minutes.” (49:01)

Timestamps for Key Segments

• 02:14 – OpenAI News: Sora, Canvas, Voice Mode with Video, Projects, Apple Intelligence/Siri
• 19:50 – AI News Beyond OpenAI/Google: iPhone SE4, Eric Schmidt, Klarna & Workforce, Cognition Devin, MSFT Phi-4
• 33:04 – Google Segment: Gemini 2.0, Agent Space, Project Astra, Android XR, Notebook LM, Project Mariner, Deep Research, AI Studio Multimodal
• 44:32 – Google Notebook LM paid tier, “Call-In” feature
• 47:20 – Project Mariner
• 48:20 – Deep Research
• 49:44 – Google AI Studio, real-time multimodal vision/voice

Listener Q&A & Tone

Jordan maintains an informal, quick-witted, and often humorous tone. He actively involves live stream comments and gives hot takes (“meh,” “nothing burger,” “bananas,” “banger”), while making technical concepts approachable for both techies and everyday professionals.

Summary Takeaway

December 16th’s episode embodies the whirlwind pace of the AI industry. OpenAI’s and Google’s race to outdo each other yields a wave of new capabilities—from Sora’s video generation and truly collaborative multimodal tools to Google’s research and agentic breakthroughs. Meanwhile, the workforce (see Klarna), user device requirements (Apple/SE4), and small model innovations (Phi-4) show AI’s tentacles reaching everywhere. Jordan’s rapid-fire insights make the news actionable—for those seeking to apply AI in business, career, or even everyday communication.

If you want to be the most AI-savvy person in the room, this week’s Everyday AI news recap is an essential listen—or, thanks to this summary, an essential read.