Cum învățăm să construim viitorul? | Andrei Stan | TEDxIasi

Cum învățăm să construim viitorul? | Andrei Stan | TEDxIasi

How do we learn to build the future? | Andrei Stan | TEDxIasi

by Andrei Stan · View the channel on PSI

Applied Sciences & Engineering Education JNEducationJNAPhilosophy and theory of education
🎙 Andrei Stan 👥 44.6M 📅 September 9, 2026 ⏱ 16 min 👁 4 📄 expert opinion 🧭 2026-09-09
Available in: English (current) Français

Keywords

artificial intelligencehardwaredomain-specific architecturesRISC-Veducation

Summary

In this TEDx talk, Andrei Stan, a professor at the Technical University ‘Gheorghe Asachi’ in Iași, addresses the question ‘How do we learn to build the future?’ He begins by defining the three pillars of modern AI: algorithms (the recipe), data (the ingredients), and hardware (the kitchen). He argues that hardware is the least visible but most decisive pillar, as even the best algorithm and data need suitable computing infrastructure to run. He traces the shift from general-purpose processors to specialized hardware, citing Hennessy and Patterson’s 2017 Turing Award lecture on domain-specific architectures and the example of GPUs, originally designed for graphics but now powering AI. He demonstrates this concept with the robot Grivai from the Spot family, which uses specialized components for perception, balance, and control. Stan then presents the educational side, emphasizing that future engineers must understand hardware-software co-design. He highlights his students’ projects and his research group’s work on a RISC-V extension that speeds up interface configuration, validated by an open-source test tool. He concludes that the real barriers are not technological but educational and creative, and calls for a holistic understanding of AI as a system combining algorithms, data, hardware, and people.

196 words

Critical Evaluation

Value of the Information & Strength of the Argument

The talk’s value lies in demystifying AI by highlighting the often-overlooked hardware foundation, and in linking technical concepts to concrete applications and pedagogical practices. The argumentation is clear and progressive: starting from the three-pillar analogy, it moves through historical context (creative destruction, Turing Award), a tangible example (GPU), a live demonstration (robot), and then to the author’s own research and student work, ultimately advocating for an integrated educational approach. The reasoning is sound and supported by recognized authorities and practical evidence, though the depth is limited by the talk format.

Scientific Rigor, Source Quality, Title Accuracy

The speaker demonstrates scientific rigour by correctly referencing Hennessy and Patterson’s Turing Award work on domain-specific architectures and Schumpeter’s concept of creative destruction. He also mentions the open-source RISC-V architecture and presents real student projects with preliminary results submitted to a conference. However, no explicit academic sources or papers are cited, and the only link provided is the general TEDx page. The title adequately reflects the content, and the talk stays on topic throughout.

179 words

Title / Content Match

The title directly reflects the content: the speaker argues that building the future requires understanding the hidden hardware pillar of AI and adapting education accordingly, and he illustrates this with examples from his own teaching and research.

Quality & Reliability

8/10

The talk provides accurate and well-structured information, referencing recognized authorities such as Hennessy and Patterson and Schumpeter, and grounds claims in concrete student projects. However, it remains a popular-science talk with no detailed citations, and the only source link is the generic TEDx page.

Key Moments

Markers derived by PSI from the transcript: the creator did not define chapters.

Cited Sources

  • TEDx official page — The video description provides this link as general information about TEDx events.

Contribution & Novelties

The talk brings a fresh perspective by focusing on the hardware pillar of AI, which is rarely discussed in popular science, and by connecting it to education and open-source hardware. It illustrates how universities in less central locations can contribute to cutting-edge R&D using open tools like RISC-V. The emphasis on teaching students to design specialized systems rather than just use existing ones is a valuable takeaway.

Pour aller plus loin :

  • RISC-V — The open-standard instruction set architecture that the speaker and his students use for their research; central to the talk’s technical aspects.
  • Graphics processing unit — The GPU as an example of a domain-specific architecture that found a new purpose in AI, as described in the talk.
  • Turing Award — The prestigious award referenced to cite Hennessy and Patterson’s contribution to computer science.
  • Creative destruction — Schumpeter’s concept invoked to explain technological paradigm shifts.

147 words

Radar Profile

The radar profile suggests a talk with moderate information density and technical depth, but high informational quality and good overall reliability. It is stronger on qualitative aspects and argumentation than on quantitative content or formal complexity.

Reliability 7/10