Beyond the Hype: What AI Actually Can (and Can't) Do • Jodie Burchell & Michelle Frost • GOTO 2026

Beyond the Hype: What AI Actually Can (and Can't) Do • Jodie Burchell & Michelle Frost • GOTO 2026

by GOTO Conferences

Humanities, Social Sciences & Thought AI — Models & Research UYQArtificial intelligence (AI)UYQFGenerative artificial intelligence
🎙 GOTO Conferences 👥 1.1M 📅 March 16, 2026 ⏱ 29 min 👁 2K 📄 expert opinion 🧭 2026-08-02
Available in: English (current) Français

Keywords

AImachine learningethicsproductivityLLM

Summary

In this GOTO Unscripted interview, Michelle Frost and Jodie Burchell, both developer advocates at JetBrains, engage in a candid conversation about the current state of AI. They begin by tracing Jodie’s unconventional career path from clinical psychology to biostatistics and NLP, highlighting the importance of statistical rigor and scientific method. The discussion then delves into the contested definition of AI, distinguishing between generative AI and classical machine learning. They emphasize that generative AI is still unproven and often conflated with more reliable classical techniques. The conversation shifts to the necessity of foundational machine learning knowledge, using a RAG system as an example to illustrate that data quality and domain expertise remain crucial. They also touch on AI ethics, noting the mathematical complexity of fairness definitions and the socio-technical challenges. The speakers reflect on the recent AI bubble discourse, sparked by GPT-5’s announcement, and draw historical parallels to previous AI summers. They conclude by examining research on developer productivity, suggesting that evidence on AI’s impact is mixed and often overstated. Throughout, they advocate for a measured, evidence-based approach to AI adoption.

180 words

Critical Evaluation

The interview provides a valuable, measured perspective on AI, grounded in the speakers’ extensive experience. Jodie Burchell’s background in clinical psychology and biostatistics lends credibility to her emphasis on scientific rigor and measurement. Michelle Frost’s expertise in AI ethics and fairness adds depth to the discussion of bias and governance. The conversation successfully demystifies AI by distinguishing between generative AI and classical machine learning, a crucial clarification often missing in public discourse. The argument for retaining foundational ML knowledge is well-articulated, using a concrete RAG example to demonstrate that data quality and domain expertise are indispensable. The discussion on AI ethics is nuanced, acknowledging the mathematical complexity of fairness and the socio-technical nature of the problem. The historical parallels to AI bubbles are insightful, providing context for current hype. However, the interview is largely opinion-based, with few direct citations to specific studies or papers. While the speakers reference research on developer productivity, they do not provide specific data or sources, limiting the ability to verify claims. The conversation is engaging and accessible, but it may lack the depth required for a purely scientific audience. The adéquation between title and content is strong, as the discussion indeed explores AI’s capabilities and limitations. Overall, the interview offers a thoughtful, expert perspective, but its reliance on anecdotal evidence and lack of systematic citations slightly diminish its scientific rigor.

225 words

Title / Content Match

The title accurately reflects the content: a measured discussion on AI's capabilities and limitations.

Quality & Reliability

8/10

The speakers are experienced data scientists and AI advocates with backgrounds in psychology, biostatistics, and AI ethics. They reference research and provide nuanced perspectives, but the discussion is largely opinion-based and lacks systematic citations.

Chapters

Cited Sources

Concurring Sources

External References

Contribution & Novelties

The interview offers a refreshingly balanced perspective on AI, emphasizing the importance of distinguishing between generative AI and classical machine learning. It highlights the enduring relevance of foundational ML knowledge and the socio-technical complexities of AI ethics. The historical context of AI bubbles provides a useful framework for understanding current hype.

Pour aller plus loin :

  • AI winter — Historical periods of reduced funding and interest in AI, relevant to the discussion of AI bubbles.
  • Transformer architecture — The foundational architecture behind most generative AI models, central to the conversation.
  • Fairness in machine learning — Overview of fairness definitions and challenges, directly related to Michelle’s expertise.
  • Retrieval-augmented generation (RAG) — The technique discussed as an example of a seemingly simple AI application that still requires careful data handling.

128 words

Radar Profile

The radar profile shows strong scores in quality of information and reliability, reflecting the speakers' expertise and measured tone. The quantity of information is moderate, and the technical level is accessible, making it suitable for a broad audience. The overall balance suggests a trustworthy but not deeply technical discussion.

Reliability 7/10