Knowledgebase Ninjas

The episode contains some substantive points about AI-ready documentation structure, the importance of domain knowledge in validation, and future trends in documentation (shorter chunks, in-product help). However, substantial portions are filler - rambling personal anecdotes about LinkedIn visibility, self-promotion advice, and vague motivational statements ('keep exploring'). The core insight that technical writers should focus on accuracy validation and chunking is valuable but takes 12 minutes to extract from 15 minutes of content.

“For that you need to have a good domain knowledge. You need to be well versed with the product that you are handling. So those are information which AI cannot have it on its own.”

“It should be chunked, it should be split into multiple sections. It's like the limited information, but that limited information should be accurate and it should be crisp.”

The core argument - that human expertise is needed to validate AI outputs and that technical writers should focus on chunked, in-product documentation - is sensible but represents mainstream thinking in the AI+docs space circa 2024. The guest recycles familiar frameworks (human-in-the-loop, shorter attention spans, structured data for AI consumption) without contrarian edge, novel methodology, or first-principles challenge to conventional wisdom.

“the human in the loop aspect that you always have to factor in”

“in the future even like people will have even less attention Spanish”

Steev Kundukulangara is a Senior Technical Content Strategist at Teradata, a legitimate database/analytics company, with genuine domain experience spanning technical support, systems analysis, and technical writing. This is a real practitioner, not a pure consultant or thought-leader. However, his visible impact is constrained to conference attendance and LinkedIn - no evidence of shipping large-scale documentation systems, leading teams, or measurable outcomes that would indicate top-tier operator status.

“Senior technical content strategist at Teradata”

“I am working like I have implemented kind of AI agents for writing, for reviewing my documentation”

The episode lacks named examples, metrics, and concrete data. The guest mentions implementing AI agents for documentation review and compressing a style guide from many lines to five-six lines, but provides no company names (other than Teradata), no performance benchmarks, no before/after comparisons, and no timeline. The advice is largely abstract (chunk documentation, validate accuracy, focus on clarity).

“I initially dumped all my style guide into the AI and it was hallucinating and it was giving all kinds of rubbish answers. But then I compressed that old MD information into like five to six lines”

“The amount of technical writers that was needed to work on a particular product has reduced significantly, like with the help of AI”

The host (Speaker B/Gauri) asks opening questions competently but rarely pushes back, challenges claims, or pursues specific follow-ups. When the guest makes broad assertions (e.g., 'organizations that are laying off technical writers completely...will see what kind of bad decision it was'), the host does not ask for evidence. The 'rapid fire' section devolves into soft softball questions ('one word that comes to mind') rather than probing deeper. There is minimal tension or productive disagreement.

“So today's documentation, uh, do you think it won't survive the next three to five years?”

“Got it, Got it. Now, uh, with AI, everybody can produce content much more easily, right?”