We obsess over what AI can automate but I keep coming back to what it can’t. Code, artifacts, yes. 6 stakeholders truly aligning, still hard. Maybe that’s the real constraint
Adi Polak
@adipolak.bsky.social
Code | Test | Deploy | Get paged Data streaming, Machine Learning and AI app intensive Systems
We say AI boosts productivity but it’s not symmetric. Engineers: 3–4x. PMs: flat. Now one PM sits on 15–20 engineers worth of output. Is this productivity… or just shifting the coordination bottleneck?
there’s something fascinating happening where developers use ai all day, but still won’t fully rely on it. that gap feels important. like we’re watching trust being earned in real time, not declared in benchmarks.
I keep wondering how many teams are racing to build custom JIRA, custom Workday, and custom CRM replacements, only to eventually ask: “When did we become an internal software company?”
5/5 ➡️ Achieving Precision in AI: Retrieving the Right Data Using AI Agents by Adi Polak bit.ly/47XWDS8 @adipolak.bsky.social #AIagents
Achieving Precision in AI: Retrieving the Right Data Using AI Agents
Adi Polak explains the path from GenAI prototype to production by focusing on precision - the competitive edge. She details Agentic RAG architectures, emergent agent design patterns, and crucial feedb...
bit.ly
the debate rages: is vibe coding good or bad? reality check-top Silicon Valley companies are already using it and seeing results. to win in the age of AI coding, think like a head chef: you don’t chop every vegetable, you design the dish. i wrote about it, check it out
in case you've missed it, here are some G𝗼𝗹𝗱-𝘀𝘁𝗮𝗻𝗱𝗮𝗿𝗱𝘀 𝘁𝗼𝗼𝗹𝘀 for AI pipelines: A/B testing randomized controlled trials natural experiments. as well as 𝘱𝘰𝘴𝘵-𝘥𝘦𝘱𝘭𝘰𝘺𝘮𝘦𝘯𝘵 𝘮𝘰𝘯𝘪𝘵𝘰𝘳𝘪𝘯𝘨
Unlock precision in GenAI by moving beyond traditional #RAG ⇨ #AgenticRAG. In this #InfoQ video, @adipolak.bsky.social dives into Agentic RAG architectures, emergent agent design patterns, and the crucial feedback loops (including LLM-as-a-judge) that drive continuous refinement. ▶️ bit.ly/47XWDS8
Amazon is laying off 30,000 corporate employees. the biggest cut in tech this year. If you know someone affected, check in. “How are you holding up?” goes a long way. 💬 Be good.
Using ChatGPT as a solution to a debate is never a good idea
Following up on my Data Streaming Patterns talk, I’ll be back for Current in New Orleans on October 29th in a few weeks. You know I got a FREE ticket code. Use CMP-ADI for free ticket and hopefully I'll see you there.
Maybe AI is a bubble. Maybe it’s the next electricity. Either way, the worst career move is sitting it out.
What advice would you give to someone in their twenties who wants to build a secure tech career?
AI alliances be like: Keep your friends close, and your cloud providers closer☁️
What are some reasons not to buy iPhone 17? Asking for a friend.
We can get so caught up in imagining the future that we forget to look for inspiration in stories from the past. It's hard to make anything truly innovative. History has some harsh lessons for us.
Software engineering is here to stay. Companies can’t replace expertise with AI.
I’ve been studying some new organizational leadership approaches in the new “AI world” and this made me pause: What if anyone at any level in an organization had an idea, and there were no barriers to bringing it to life?
Cloud let you deploy in minutes. AI lets you build features that think. Learn to compose AI like services, and you’ll own the next wave of apps.
Ambition is refusing to quit on ourselves. Leadership is refusing to quit on others. Let it sink.
You might already be skilled in ETL scheduling, analytics queries, or machine learning integration—but are you ready to support AI agents? In her 🆕 article in @thenewstack.io, @adipolak.bsky.social dives into the skills data engineers need in the age of AI! ⤵️
Data Engineering in the Age of AI: Skills To Master Now
A look at the critical capabilities data engineers must develop to stay relevant and valuable, as well as practical ways to sharpen those skills.
thenewstack.io
Have LLMs actually made peoples coding better or just faster?