Practical AI · Practical AI LLC

Tool calling and agents

February 14, 2025·45 min·4 clips
Daniel reveals why asking an LLM to create a HubSpot deal is like nails on a chalkboard for AI experts.
Practical AI hosts Daniel Whitenack and Chris Benson catch up on a week dense with AI news in their fully connected format. The dominant story is Elon Musk's public bid to acquire OpenAI, which CEO Sam Altman rejected publicly with the statement that OpenAI is not for sale. The hosts note that Musk and Altman's relationship — former collaborators turned adversaries — has been a source of ongoing drama, and treat the acquisition bid as an extension of that personal rivalry rather than a serious strategic proposition. The OpenAI Super Bowl commercial, which reportedly cost around $14 million, is discussed as a notable moment: the ad featured the animated dots associated with the ChatGPT interface forming scenes, building to recognition of the brand. The hosts note that $14 million represents a small fraction of OpenAI's infrastructure and compute costs, contextualising it as a brand-awareness investment rather than an unusual expenditure. The broader question of OpenAI's financial trajectory — continuing to lose enormous sums on model hosting and compute while growing revenue — is raised as a structural issue that neither the Musk bid nor the Super Bowl ad resolves. The hosts reflect on whether tracking billionaire AI feuds is a meaningful signal about the direction of the industry or primarily entertainment, and both conclude it falls more in the entertainment category. Other AI news from the week is briefly covered, with the hosts noting the pace of development makes comprehensive coverage impossible. The episode is conversational and opinion-forward, giving regular Practical AI listeners a weekly orientation to the AI news landscape.

As heard by us

A practical reset on what agents do, and what tool calling actually is.

The episode treats tool calling less like a trick and more like a practical boundary between the model and the system around it. Daniel and Chris keep the focus on how an LLM actually connects to another system, so the conversation stays concrete instead of drifting into jargon.

Read the full review in PlayNext →

Why you'd press play

Stop mixing up agents, tools, and LLMs.

Read the full recommendation in PlayNext →
Listen to the show on