Nadia Dubois

Nadia Dubois treats AI systems less as abstractions and more as products to be benchmarked, priced, and stress‑tested in real workloads. At the masthead she covers technology, but in practice she zeroes in on large language models, developer tools, and cloud AI platforms, with a focus on how they perform, what they cost, and which one is the better buy for a given use case.

Side‑by‑side testing of frontier AI models

A defining strand of Dubois’s work is structured comparison of competing AI models under realistic constraints. In her coverage of Gemini and ChatGPT, for example, she builds tests that combine quality, latency, and token limits with a close reading of API pricing, turning abstract model specs into concrete trade‑offs for engineers and product teams. That approach carries through other comparison pieces where she runs the same prompts, coding tasks, or document‑analysis jobs across multiple models and reports out both strengths and failure modes.

She writes these tests in clear, operational language, grounding her judgments in measurable outcomes rather than hype. Instead of broad claims about “best” models, she breaks the problem down: which API handles long‑context workloads, which one gives cleaner code on first pass, where safety filters become a practical blocker, how rate limits behave under load. The result is coverage that reads like a lab report adapted for a technical but time‑pressed audience, with enough detail for specialists and enough framing for decision‑makers.

Cost, tokens, and the economics of AI adoption

Dubois repeatedly returns to the question of cost, especially where token pricing and context windows shape what is possible to build.

She dissects per‑million‑token rates, input vs output pricing, and the sometimes non‑obvious ways that context length and compression affect total spend on a project. Her stories often hinge on specific, quantified gaps between APIs, and she shows how ostensibly small price differences compound at the scale of high‑volume applications or million‑token contexts.

This emphasis on economics distinguishes her work from more purely capability‑driven AI reporting. She writes for readers who need to balance performance, reliability, and budget, so she pairs benchmarks with simple scenario math: how much a feature will cost to run per user, where a model’s cheaper rate offsets slightly weaker quality, and when premium models justify their price in reduced development time.

That framing makes her AI coverage particularly relevant for founders, engineering leaders, and teams under cost pressure.

Hands‑on developer workflows and tooling

Another constant in Dubois’s beat is the developer experience around AI, not just the models themselves. She spends time on API ergonomics, SDKs, context‑management tools, and evaluation frameworks, often walking through the exact steps a developer would take to wire a model into an existing stack. When she writes about new features such as extended context, function calling, or multimodal input, she anchors the story in how these capabilities change the shape of real applications rather than treating them as isolated announcements.

Her pieces frequently surface practical edge cases: how streaming outputs behave, what breaks when prompts or payloads approach context limits, how rate limiting interacts with batch processing, and where vendor documentation leaves gaps. She favors concrete examples - code generation, agents, document summarization pipelines - over generic claims, and she pays attention to debugging and evaluation, not just the happy path.

This gives her work a workshop quality that aligns with how engineers actually test and adopt new tools.

Platform strategy and the shifting AI landscape

Beyond narrow benchmarking, Dubois tracks how the major AI providers reposition themselves over time. She follows changing price structures, new model tiers, and shifts in default offerings, using them to explain where the underlying platforms are headed and how that impacts lock‑in, portability, and negotiation leverage for customers. Her analysis connects product changes to broader competitive dynamics between AI labs and cloud providers, but she keeps the focus on what it means for people building on top of these systems.

She also pays attention to ecosystem moves around these platforms, including third‑party evaluation tools, guardrail products, and orchestration layers that sit between applications and raw model APIs.

Instead of treating these as separate beats, she folds them into a single story about how AI moves from demo to dependable infrastructure. That through‑line - tested performance, clear economics, and developer‑centric detail - is what sets her coverage apart on the technology desk.