How To Access GPT-4 (and what it can do)

How To Access GPT-4 (and what it can do)

by Matt Wolfe

Applied Sciences & Engineering AI — Models & Research UYQArtificial intelligence (AI)UYQFGenerative artificial intelligence
🎙 Matt Wolfe 👥 1.0M 📅 March 15, 2023 ⏱ 23 min 👁 368K 📄 news review 🧭 2026-08-28
Available in: English (current) Français

Keywords

GPT-4ChatGPT PlusOpenAIAI capabilitiesGoogle Workspace

Summary

In this video, Matt Wolfe provides a comprehensive overview of the GPT-4 release, announced by OpenAI on March 14, 2023. He begins by covering Google’s simultaneous announcement of AI features in Workspace, which was overshadowed by GPT-4 news. He then details the key capabilities of GPT-4 as outlined by OpenAI: handling over 25,000 words of text, visual inputs, and enhanced creativity. He notes that many of these features are not yet available in the ChatGPT interface, which currently offers a limited version of GPT-4 to ChatGPT Plus subscribers. The video includes a recap of OpenAI’s live demo, highlighting the impressive ability to turn a hand-drawn website mockup into functional HTML code and to apply tax code calculations. Matt then conducts his own tests, comparing GPT-4 with GPT-3.5 on tasks like writing a story and generating HTML code. He observes that GPT-4 produces longer, more descriptive outputs but is slower and sometimes cuts off mid-generation. He concludes that while GPT-4 shows significant improvements, the most anticipated features like image input and web browsing are not yet accessible to the public, and he expresses excitement for future rollouts.

186 words

Critical Evaluation

Value of the Information & Strength of the Argument

The video offers valuable, hands-on insights into GPT-4’s capabilities at launch, going beyond the official announcement by testing the model in real-time. The creator’s comparisons between GPT-4 and GPT-3.5 provide concrete examples of improvements in creativity and code generation, while also noting limitations such as speed and output truncation. The argumentation is balanced, acknowledging both the impressive potential and the current constraints. However, the analysis is largely anecdotal and based on a few specific tests, which limits its generalizability. The creator also speculates on future features without providing concrete timelines, which is understandable given the early stage of release.

Scientific Rigor, Source Quality, Title Accuracy

The video references official OpenAI materials, including the GPT-4 announcement page and a blog post with benchmarks, which adds credibility. The creator also mentions the OpenAI live stream, providing a secondary source for the demonstrations. However, the video does not cite independent research or expert opinions, and the hands-on tests are not systematically documented. The title accurately reflects the content, focusing on access and capabilities. The video is a news review, and while it is informative, it lacks the rigor of a formal technical analysis. The creator’s own tests are not peer-reviewed and may not be reproducible, but they are presented as personal observations, which is appropriate for the format.

224 words

Title / Content Match

The title accurately reflects the content: the video explains how to access GPT-4 (via ChatGPT Plus) and demonstrates its capabilities through live tests.

Quality & Reliability

7/10

The video provides a timely and practical overview of GPT-4's release, including hands-on testing and comparisons with GPT-3.5. The creator clearly distinguishes between announced features and those actually available, and includes references to official OpenAI materials. However, the analysis is largely anecdotal and lacks deep technical verification.

Key Moments

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

Cited Sources

  • FutureTools.io — Creator's website for curating AI tools, mentioned as a resource.
  • FutureTools Discord — Community Discord, mentioned for discussions and requests.
  • Matt Wolfe's Blog — Creator's personal blog.
  • Mubert — Music generation tool used for the outro music.
  • FutureTools Desktop Backgrounds — Free desktop backgrounds offered by the creator.
  • FutureTools Newsletter — Weekly newsletter with AI news and tools.

Concurring Sources

  • OpenAI GPT-4 Announcement — Official OpenAI page confirming GPT-4 features and availability.
  • GPT-4 Technical Report — Technical paper providing benchmarks and details on GPT-4.

Dissenting Sources

  • No discordant sources identified — The video's claims align with official OpenAI information and other early reviews.

Contribution & Novelties

This video provides a timely, hands-on look at GPT-4’s release, offering practical comparisons with GPT-3.5 and highlighting the gap between announced features and actual availability. It serves as a useful snapshot for early adopters, though it does not introduce novel technical insights beyond what was already in OpenAI’s official materials.

Pour aller plus loin :

  • GPT-4 Technical Report — The official technical report detailing GPT-4’s architecture and capabilities.
  • OpenAI GPT-4 Announcement — Official page with feature overview and examples.
  • Google Workspace AI Announcement — Google’s blog post on AI features in Workspace, mentioned in the video.
  • Be My Eyes — App using GPT-4 for visual accessibility, cited in the video.
  • Duolingo — Language learning platform using GPT-4 for conversational practice, cited in the video.

124 words

Radar Profile

The radar profile shows high scores in information quantity and quality, reflecting the video's comprehensive coverage and practical testing. The technical level is moderate, suitable for a general audience, and reliability is good due to alignment with official sources. The overall profile suggests a well-rounded, informative video with minor limitations in depth.

Reliability 7/10

💬 Très positif. Sur les 30 commentaires analysés, les spectateurs expriment un enthousiasme marqué pour GPT-4 et la rapidité de la couverture, avec des anecdotes personnelles sur l'utilisation de l'IA et des remerciements pour la clarté de la présentation.