# Large Language Models on We Get To Do This

Large Language Models comes up in 7 chapters across 5 episodes of We Get To Do This, from September 2025 to October 2026.

- **Chapters:** 7
- **Episodes:** 5
- **Discussed:** September 2025 – October 2026
- **Related topics:** [Artificial Intelligence](https://podherd.com/wgtdt/topic/artificial-intelligence.md), [Chatgpt](https://podherd.com/wgtdt/topic/chatgpt.md), [Perplexity](https://podherd.com/wgtdt/topic/perplexity.md), [Automation](https://podherd.com/wgtdt/topic/automation.md), [Data Centers](https://podherd.com/wgtdt/topic/data-centers.md), [Grok](https://podherd.com/wgtdt/topic/grok.md), [Marketing](https://podherd.com/wgtdt/topic/marketing.md), [Nashville](https://podherd.com/wgtdt/topic/nashville.md), [Sam Altman](https://podherd.com/wgtdt/topic/sam-altman.md), [Silicon Valley](https://podherd.com/wgtdt/topic/silicon-valley.md)

## Chapters 1–7 of 7

### Data Center Infrastructure and In-Flight AI Usage

- **Episode:** [53: The Uncomfortable Truth](https://podherd.com/wgtdt/episode/53-the-uncomfortable-truth.md)
- **Published:** 2026-10-07
- **Time:** 9:46–13:16
- **Transcript:** https://podherd.com/wgtdt/episode/53-the-uncomfortable-truth/chapter/data-center-infrastructure-and-in-flight-ai-usage.md

The discussion clarifies that data centers are essentially virtual computers running large language models, comparing them to established services like iCloud and Facebook. One host recounts a "light year speed" experience using Grok AI while flying from Nashville, realizing the power of cloud-based processing. They conclude that while AI can return hours of time to users, that time must be spent wisely rather than just filled with more work.

Keywords: [data centers](https://podherd.com/wgtdt/topic/data-centers.md), icloud, [grok](https://podherd.com/wgtdt/topic/grok.md), [large language models](https://podherd.com/wgtdt/topic/large-language-models.md), [nashville](https://podherd.com/wgtdt/topic/nashville.md), processing speed

### AI Marketing Deception and the Definition of Intelligence

- **Episode:** [43: Danger Will Robinson!](https://podherd.com/wgtdt/episode/43-danger-will-robinson.md)
- **Published:** 2026-07-29
- **Time:** 41:13–46:14
- **Transcript:** https://podherd.com/wgtdt/episode/43-danger-will-robinson/chapter/ai-marketing-deception-and-the-definition-of-intelligence.md

An analysis of the "evil" marketing strategies used by Silicon Valley to frame AI as either a savior or an existential threat to create regulatory moats. The hosts define AI as "simulated intelligence" rather than true reasoning, describing it as a sophisticated tool or "screwdriver" for data categorization. They argue that machines lack the ability to understand the nuances of the human condition or make sound moral judgments.

Keywords: [artificial intelligence](https://podherd.com/wgtdt/topic/artificial-intelligence.md), [silicon valley](https://podherd.com/wgtdt/topic/silicon-valley.md), [marketing](https://podherd.com/wgtdt/topic/marketing.md), reasoning, [large language models](https://podherd.com/wgtdt/topic/large-language-models.md), [automation](https://podherd.com/wgtdt/topic/automation.md)

### Perplexity vs ChatGPT and AI Workflow Efficiency

- **Episode:** [43: Danger Will Robinson!](https://podherd.com/wgtdt/episode/43-danger-will-robinson.md)
- **Published:** 2026-07-29
- **Time:** 51:06–55:17
- **Transcript:** https://podherd.com/wgtdt/episode/43-danger-will-robinson/chapter/perplexity-vs-chatgpt-and-ai-workflow-efficiency.md

A comparison of different AI tools like Perplexity, ChatGPT, and Gemini, noting their frequent contradictions and factual errors. While the hosts use AI to improve administrative workflows and list-making, they emphasize the need for constant human fact-checking. They conclude that AI is a useful tool for search categorization but remains subjective and often outdated.

Keywords: [perplexity](https://podherd.com/wgtdt/topic/perplexity.md), [chatgpt](https://podherd.com/wgtdt/topic/chatgpt.md), gemini, workflows, fact-checking, [large language models](https://podherd.com/wgtdt/topic/large-language-models.md)

### AI as a Tool for Craftsmen and Final Sign-off

- **Episode:** [29: Messiah Meme](https://podherd.com/wgtdt/episode/29-messiah-meme.md)
- **Published:** 2026-04-22
- **Time:** 54:23–57:00
- **Transcript:** https://podherd.com/wgtdt/episode/29-messiah-meme/chapter/ai-as-a-tool-for-craftsmen-and-final-sign-off.md

The hosts conclude by framing AI as a mechanical tool rather than a relational entity, citing Exodus 31 regarding the spiritual gift of craftsmanship. One host shares an anecdote about his "robot" assistant attempting to act human by suggesting a bedtime, which he rejected to maintain clear boundaries between man and machine. The episode ends with a blessing and a reminder of the gift of life.

Keywords: exodus, craftsmanship, [large language models](https://podherd.com/wgtdt/topic/large-language-models.md), [chatgpt](https://podherd.com/wgtdt/topic/chatgpt.md), imago dei

### AI Expectations and Grok 4.1 Release

- **Episode:** [11: Hair on Fire](https://podherd.com/wgtdt/episode/11-hair-on-fire.md)
- **Published:** 2025-11-19
- **Time:** 15:25–18:15
- **Transcript:** https://podherd.com/wgtdt/episode/11-hair-on-fire/chapter/ai-expectations-and-grok-4-1-release.md

The rapid advancement of artificial intelligence, specifically the release of Grok 4.1, has created an expectation for instantaneous results that contrasts with natural processes like heating a home or cooking. While Grok 4.1 promises improved emotional and collaborative interactions, there are concerns about the long-term financial viability of the AI market. Future church operations may include private, decentralized large language models to avoid dependence on external providers.

Keywords: [artificial intelligence](https://podherd.com/wgtdt/topic/artificial-intelligence.md), grok 4.1, hvac, [large language models](https://podherd.com/wgtdt/topic/large-language-models.md), xai

### AI Religious Overtones and the "Godfather of AI"

- **Episode:** [2: Tall Cotton](https://podherd.com/wgtdt/episode/2-tall-cotton.md)
- **Published:** 2025-09-10
- **Time:** 28:34–30:32
- **Transcript:** https://podherd.com/wgtdt/episode/2-tall-cotton/chapter/ai-religious-overtones-and-the-godfather-of-ai.md

A news report triggers a discussion on the religious language used by AI developers like Jeffrey Hinton and Sam Altman. The hosts argue that AI is a "parlor trick" or a fast copy-paste engine rather than true intelligence. They suggest that society has been programmed to accept talking computers through pop culture icons like "HAL" and "Knight Rider."

Keywords: jeffrey hinton, [sam altman](https://podherd.com/wgtdt/topic/sam-altman.md), [artificial intelligence](https://podherd.com/wgtdt/topic/artificial-intelligence.md), [large language models](https://podherd.com/wgtdt/topic/large-language-models.md), lost in space

### AI Scraping, Model Collapse, and the Hegelian Dialectic

- **Episode:** [2: Tall Cotton](https://podherd.com/wgtdt/episode/2-tall-cotton.md)
- **Published:** 2025-09-10
- **Time:** 40:16–44:17
- **Transcript:** https://podherd.com/wgtdt/episode/2-tall-cotton/chapter/ai-scraping-model-collapse-and-the-hegelian-dialectic.md

The technical limitations of AI are explored, specifically how Large Language Models (LLMs) scrape existing data to create "copies of copies." The hosts compare this to the Hegelian Dialectic and the third law of thermodynamics (entropy), suggesting that AI output will eventually lose resolution and fidelity. They warn that this process can lead to the degradation of truth and original thought.

Keywords: [perplexity](https://podherd.com/wgtdt/topic/perplexity.md), large language model, hegelian dialectic, entropy, model collapse

---

This is the markdown version of a page on We Get To Do This for AI agents. Every public page on this site has one: add `.md` to its URL, or request it with `Accept: text/markdown`.

- [Site index for agents (llms.txt)](https://podherd.com/wgtdt/llms.txt)
- [Home](https://podherd.com/wgtdt/index.md)
- [Search transcripts](https://podherd.com/wgtdt/search.md)
- [All episodes](https://podherd.com/wgtdt/episodes.md)
- [Topics](https://podherd.com/wgtdt/topics.md)
- [People](https://podherd.com/wgtdt/people.md)
