Navaid Abidi

While watching Claude code and writing a solution comparison during a PoC, I figured it might be handy to have a cost comparison for the client. Copied the terminal output into Claude Coworker, and it produced perfect slides in PPT. The integration of formerly disjunct silos across the product development process eliminates coordination and information costs, making it much more effective.

8

💡 AI “ARR” vs. Real Revenue Lately, we’ve seen AI startups announce eye-popping “ARR” numbers. But dig deeper and you realize a lot of it isn’t true software revenue at all — it’s GMV (Gross Merchandise Value) passing through from LLM providers like OpenAI. GMV is a great signal of activity, but it’s not the same as the net revenue a company actually keeps. If you confuse the two, you end up with inflated growth stories that hide razor-thin margins. That’s why this passage stuck with me — it nails the difference in a way that’s both sharp and memorable: Feed Your Dog, and Your Startup Passing through a big number is tasty… lots of calories, not much nutrition. Top-line GMV can be a useful leading indicator. But if you’re using it as ARR, you’re feeding your startup an artificial revenue diet. When you look at it this way, the “ARR” many AI startups tout is really GMV dependent on someone else’s GMV coming through. It blurs the line between service and product and raises the real question: How recurring, and how valuable, is that revenue, once you strip out the pass-through? And how valuable is it once you realize it’s predicated on someone else’s pass through? And if you don’t feed your dog (and your startup) real net revenue… well, you know how that ends. You don’t have a dog.

12

1 Comment

🔍 Why the Future of AI Research Isn’t Just About Retrieval—It’s About Deep Reasoning Large Language Models (LLMs) are no longer satisfied with simply fetching information. Enter Deep Research (DR) Agents—a new class of autonomous AI designed to tackle complex, multi-step research workflows that blend: ✅ Dynamic Reasoning (thinking step by step) ✅ Adaptive Planning (adjusting strategies on the fly) ✅ Tool Use (coding, analyzing data, browsing the web) ✅ Structured Report Generation Unlike traditional retrieval-augmented generation (RAG), DR agents don’t just retrieve and repeat—they synthesize, validate, and plan. What sets them apart? • Hybrid Search: Combining API calls with live browser exploration to handle both structured and messy, real-time content. • Multi-Agent Collaboration: Specialized agents work together, mimicking research teams. • Non-Parametric Learning: They evolve capabilities without retraining, learning by storing and reusing prior workflows. • Dynamic Workflows: From clarifying your intent to planning and executing tasks, all in real time. This paradigm shift has big implications: 🏥 Healthcare: Faster synthesis of the latest studies. 💼 Finance: Real-time risk assessments with live market data. 🧪 Scientific Discovery: Automated literature reviews and hypothesis generation. The paper also highlights key challenges—like aligning evaluation benchmarks with real-world needs and optimizing multi-agent coordination. How soon do you see DR agents becoming your co-pilot in complex decision-making? #AI #DeepResearchAgents #LLMs #FutureOfWork #AIResearch

41

12 Comments

🧠'Agent Learning via Early Experience' As an AI Architect, this work stands out for its pioneering approach to training language agents—moving beyond passive imitation learning and bridging toward a true era of experience. Instead of relying on expert demonstrations or reward-based RL, the authors introduce a scalable middle-ground: agents learn by interacting with environments, then using those resulting states as direct supervision, without reward signals. 🔍 Key insights: • Motivation: Real-world agents often face hard-to-scale expert data and sparse—or absent—rewards. This paradigm lets agents learn from their own actions, scaling up diversity and generalization. • Methods: Two main strategies—implicit world modeling (internalizing environment dynamics) and self-reflection (learning from 'what went wrong'). Both use agents' own experiences to improve reasoning and robustness. • Results: Evaluated across 8 environments and major model families (Llama, Qwen), these approaches consistently outperform standard imitation learning, especially in complex and out-of-domain tasks. Early experience even boosts downstream RL, serving as a strong warm start for further finetuning. • Implications: Practical signal for agentic AI system design—build architectures that leverage interaction-derived learning, not just rewards or pre-collected demos. This can make agents more resilient, adaptive, and capable in open-ended domains (web, tools, APIs). Curious how 'early experience' could transform future agent frameworks and LLM deployments? 🔗 Attaching the original research PDF for those who want to dive deeper! #AI #LLM #AgenticAI #ReinforcementLearning #LanguageAgents #AIArchitecture

10

Integrating large language models and powerful algorithms into your workflows is a massive undertaking. All that hard work can be undermined if the outputs fall short, quickly eroding stakeholder interest and trust. That's why effectively evaluating LLMs and algorithms is crucial. From initial prototypes to production, and through continuous experimentation and benchmarking, mastering evaluation ensures these advanced tools truly deliver on their promise. #LLM #Algorithms #AI

8

This section on how dbt Labs transitioned from a purely PLG motion to layering on enterprise is a fascinating one. Two instructive passages that I have bolded (below from First Round Review's path to PMF series featuring dbt Labs) TLDR: GTM comprises your ICP, channel, and message. When you transition from PLG / bottom up to Enterprise / top down motion, naturally your ICP and channel also changes, but your messaging / proposition needs to change too, e.g., the enterprise may be multiple personas and are buying assurance as much as solution. --- "Handy found product-market fit organically for dbt as an open-source tool mostly used by data practitioners and developers. But a few years into running a commercial business, he realized he had to build a growth curve all over again with C-suite data leaders. “Even though we had an unbelievable amount of market pull, as we initially commercialized, it wasn’t easy for us to transform this open-source command line tool into a product that enterprises would pay a million dollars for,” says Handy. “When you have enough product-market fit, sometimes it allows you to get away with not being super tight on product marketing or sales motions. So around 2022, we went from this gigantic acceleration curve and overnight we realized, we have to sit at the adults’ table and figure things out real fast,” says Handy. After the PLG flame started to fizzle, Handy turned his attention to layering on a sales-led motion for the cloud platform. “We had to focus our efforts on telling cohesive stories to senior data leaders. We had to have a very clear, explainable answer to the question, ‘Why should I use the commercial product and not the open-source product? And it had to be digestible by someone with a C in their title,” he says. Handy’s answer: dbt Cloud can handle complex data for companies of every size. “The longer people used dbt, the more complex their code became,” he says. “It was a problem for the most sophisticated dbt users, who were often at the largest companies. So there was a real opportunity for us to step in and solve that for them with dbt Cloud.” To tell that story to enterprise customers, Handy relied on data, naturally. “At a user conference we presented a chart that showed the number of dbt projects that had a certain number of models in them — over 100, over 1,000, et cetera,” he says. “We watched that number climb and we knew as users ourselves, ‘Oh my God, trying to work in a dbt project with 5,000 models in it is challenging.’ So we started with that quantitative data point and asked folks in our community about their experiences with these very large, complex dbt projects, and validated that this was a pain in the ass without a cloud platform.”

23

4 Comments

𝗧𝗵𝗲 𝗵𝗶𝗱𝗱𝗲𝗻 𝗰𝗼𝘀𝘁 𝗼𝗳 𝗔𝗜 𝗮𝘁 𝘀𝗰𝗮𝗹𝗲. He was the architect of their intelligent document processing platform. Built on Vertex AI. Fine-tuned transformers. Auto-scaling via Cloud Run. 2 million documents a month across 5 languages. It worked like magic. The Tech-Director called it a breakthrough. The sales deck showcased it on every slide. Customers loved it. Until the CFO saw the cloud bill. ₹𝟭.𝟮 𝗖𝗿𝗼𝗿𝗲𝘀/𝗺𝗼𝗻𝘁𝗵 𝗶𝗻 𝗚𝗖𝗣 𝘀𝗽𝗲𝗻𝗱. → ₹85L on inference → ₹20L on pipelines → ₹15L on storage, logging, and APIs 𝗥𝗲𝘃𝗲𝗻𝘂𝗲 𝗳𝗿𝗼𝗺 𝘁𝗵𝗲 𝗔𝗜 𝗳𝗲𝗮𝘁𝘂𝗿𝗲: ₹𝟭𝟭 𝗟𝗮𝗸𝗵𝘀/𝗺𝗼𝗻𝘁𝗵. Suddenly, the ask wasn’t “How can we make this better?” It was: “Can we afford to keep this alive?” So began his second act - not as an innovator, but as a surgeon.  • Rewrote pipelines in Cloud Functions to cut Vertex AI usage  • Switched from online to batch inference  • Distilled the model to a smaller size  • Lowered OCR quality settings to save credits  • Tradeoffs everywhere - accuracy, latency, user experience It worked. The cloud bill dropped by 78%. The feature survived - barely. But the “magic” slowly disappeared. And here’s the hard truth: In cloud AI, performance is rented. Intelligence is cheap to prototype, but expensive to scale. And the best engineers? Not just the ones who build powerful systems - but the ones who know what to sacrifice when power becomes too costly. How do you write “𝘴𝘢𝘤𝘳𝘪𝘧𝘪𝘤𝘦𝘥 𝘢𝘤𝘤𝘶𝘳𝘢𝘤𝘺 𝘵𝘰 𝘴𝘵𝘢𝘺 𝘢𝘧𝘭𝘰𝘢𝘵” on your resume? In the GenAI boom, let’s not forget: Innovation is exciting but 𝘴𝘶𝘴𝘵𝘢𝘪𝘯𝘢𝘣𝘪𝘭𝘪𝘵𝘺 𝘪𝘴 𝘴𝘶𝘳𝘷𝘪𝘷𝘢𝘭. #AI #CloudComputing #GenAI #MLOps #VertexAI #AIEngineering #TechLeadership #FinOps #ProductReality #CloudCosts

23

5 Comments

Single-method retrieval often fails in #financial and #regulatory domains, exact-term search misses semantic relationships, while dense models overlook structured identifiers. A multi-path approach resolves this gap. The Multi-Path Retrieval (#MPR) pipeline implements four complementary retrieval strategies with a domain-specialized re-ranking mechanism to balance lexical precision and semantic understanding: #BM25 Sparse Retriever – captures exact matches for regulatory codes, ticker symbols, and key phrases. #Dense Semantic Retriever – identifies paraphrases and conceptually similar expressions through high-dimensional embeddings. #Metadata-Aware Retriever – filters at section-level granularity, narrowing retrieval to relevant document parts before detailed chunk search. #HyDE Retriever – generates a hypothetical answer, embeds it, and retrieves real text segments with similar structure and terminology, bridging the question–answer linguistic gap. Final results can be aggregated through weighted rank fusion (like semantic > BM25 > metadata > HyDE ), producing top-ranked chunks enriched through context bundling when adjacent similarities exceed 0.85. This ensures coherent, context-complete passages are provided to downstream models. Such architecture is especially effective in financial compliance and regulatory analysis, where both formal precision and contextual interpretation are essential. The hybrid fusion enhances #recall, #contextualaccuracy, and #interpretability in retrieval-augmented systems(#RAG). #GenAI #Banking #AIEngineer #LLM #Fraud

4