AI Tools: UNDERRATED vs OVERRATED

Key Concepts

• Large Language Models (LLMs): AI models designed to understand and generate human-like text. (GPT, Deepseek, Manis, Open Claw, Clawed, Grock, Gemini)
• Model Rating/Underrating: Assessment of an LLM’s capabilities relative to its current perception and usage.
• Cost-Effectiveness: The balance between an LLM’s performance and its associated expense.
• Integration: The seamless incorporation of an LLM into existing products and workflows.
• Creativity (in AI): An LLM’s ability to generate novel and imaginative outputs (text, images, videos).

---

LLM Performance & Current Perception

The core argument presented is a re-evaluation of the current perception of several prominent Large Language Models (LLMs). The speaker contends that GPT is “wildly overrated” and is being used as a default option despite not being at the forefront of model capabilities. This suggests a reliance on brand recognition over objective performance.

Deepseek: A Powerful Contender

Deepseek is positioned as significantly “underrated”, described as a “powerful model” actively competing with leading AI developers. The emphasis is on its potential and current capabilities, implying it deserves greater attention and usage. No specific performance metrics are provided, but the statement highlights its competitive standing against “the smartest people in the world” developing AI.

Manis: Power vs. Price

Manis is also considered “slightly underrated” due to its substantial power. However, a key criticism is its cost. The speaker believes Manis is “more expensive than it should be”, suggesting a barrier to wider adoption despite its capabilities. This introduces the concept of cost-effectiveness as a crucial factor in LLM evaluation.

Open Claw: The "Killer App"

Open Claw receives strong praise, being labeled the “killer app of AI” and therefore “underrated”. This implies it offers a uniquely valuable and impactful application of AI technology. The specific application isn’t detailed, but the strong language suggests a transformative potential.

Clawed & Opus 46: Appropriately Rated & Powerful

Clawed is deemed “properly rated”, indicating its current perception aligns with its actual performance. Specifically, Opus 46 (presumably a version or component of Clawed) is described as “wild”, signifying exceptional power and capability. This provides a concrete example supporting the overall assessment of Clawed.

Grock: Underrated Creativity

Grock is identified as “underrated” specifically for its creative output. The speaker highlights its ability to generate “videos and photos” alongside text, emphasizing its versatility and imaginative potential. This points to a strength in multimodal AI generation.

Gemini: Appropriately Rated & Underused Integration

Gemini is considered “appropriately rated” but also “underused”. The speaker emphasizes the value of its integration “in every product”, noting that many users are unaware of this widespread implementation. This suggests Gemini’s strength lies in its accessibility and seamless integration rather than necessarily groundbreaking individual performance. The speaker states they “love” this integration.

Logical Connections & Synthesis

The transcript presents a comparative assessment of several LLMs, moving from a critique of overhyped models (GPT) to highlighting undervalued ones (Deepseek, Manis, Open Claw, Grock). The assessments aren’t based on detailed benchmarks but rather on the speaker’s perception of their capabilities and value. A recurring theme is the importance of looking beyond brand recognition and considering factors like cost-effectiveness, integration, and specific strengths (like creativity). The overall takeaway is a call for a more nuanced evaluation of LLMs, recognizing that the most popular option isn’t always the best, and that several powerful models are currently flying under the radar.