Nick Ciubotariu
Seattle, Washington, United States
19K followers
500+ connections
About
I’m a tech leader, curious builder, and unrelenting tinkerer. Over two decades, I’ve…
Articles by Nick
-
Syntax Is Dead. Long Live Human Engineering
Syntax Is Dead. Long Live Human Engineering
*** written by an actual human *** Software engineers are not being replaced by AI. Stop saying and believing that.
3 Comments -
The Inevitability of Software-Defined Manufacturing
The Inevitability of Software-Defined Manufacturing
From 2003 to 2006, I worked at a contract manufacturing company as a robotics engineer. I was the first software…
3 Comments -
Life lessons from two complete strangers
Life lessons from two complete strangers
Disclaimer and proviso: "The postings on this site are my own and don't represent Amazon's position in any way…
38 Comments -
Promote with pragmatism: a rational, logical, transparent (and stress-free!) promo process
Promote with pragmatism: a rational, logical, transparent (and stress-free!) promo process
Disclaimer and proviso: "The postings on this site are my own and don't represent Amazon's position in any way…
56 Comments -
Hacking corporate jargon: the "engineering culture" and what that really means
Hacking corporate jargon: the "engineering culture" and what that really means
Disclaimer and proviso: “The postings on this site are my own and don’t represent Amazon’s position in any way…
68 Comments -
An Amazonian's response to "Inside Amazon: Wrestling Big Ideas in a Bruising Workplace"
An Amazonian's response to "Inside Amazon: Wrestling Big Ideas in a Bruising Workplace"
Disclaimer and proviso: "The postings on this site are my own and don't represent Amazon's position in any way…
1,178 Comments -
Shred the first envelope
Shred the first envelope
Disclaimer and proviso: "The postings on this site are my own and don't represent Amazon's position in any way…
5 Comments -
"The Work Death March" - a survivor's tale
"The Work Death March" - a survivor's tale
4/3/2016. Update: I received a call today, from an Executive that was close to the situation at the time, taking…
75 Comments -
Ace the coding interview, every time
Ace the coding interview, every time
Disclaimer and proviso: "The postings on this site are my own and don't represent Amazon's position in any way…
134 Comments -
Get to "Yes!" - Closing the perfect candidate
Get to "Yes!" - Closing the perfect candidate
Disclaimer and proviso: "The postings on this site are my own and don't represent Amazon's position in any way…
5 Comments
Activity
19K followers
-
Nick Ciubotariu reposted thisNick Ciubotariu reposted thisUPDATE: Figma reached out and issued a full refund ($2,520) for the inactive seats — appreciate the quick response and resolution. They also mentioned they’re working on improvements to the seat model — curious to see how that rolls out in practice. ----- Figma YOU HAVE SOME EXPLAINING TO DO. I found out I’ve been paying for multiple seats for OVER A YEAR. Apparently, if someone requests edit access to your file, and you grant it - FIGMA ADDS IT AS A PAID SEAT TO YOUR ACCOUNT. Which is absolutely crazy because that person is already paying for their OWN Figma subscription. Was it silly of me to assume it works like Google Docs? Maybe. BUT they don’t ask. They don’t notify you. There is no confirmation. They just add another user to your subscription. This is straight up predatory practice. I did some digging and I am far from the first person this has happened to. So. Apparently I've been paying for 8 seats this whole time. Each seat costs $16/month. I went back through all my invoices, and I've overpaid by around $2,500. Could I have been checking my statements more closely? Maybe, but you know how it is. We've all paid for random subscriptions without realizing. (Gym membership, I'm looking at you.) Best part? When they (finally) answered me ten minutes ago, they gave me a refund of $150. On a $2,500 overcharge. So this is me offering (free) advice: Guys, check your statements. Seriously. And more importantly, please reshare this and tag Figma. I want this loud enough that they actually deal with it.
-
Nick Ciubotariu reposted thisNick Ciubotariu reposted thisBecoming a great AI-first software engineer takes 10 years and one day. 10 years to become a great engineer. One day to learn the coding agent. Don't skip the first part.
-
Nick Ciubotariu posted thisI'm blown away by Opus 4.7. I started working with AI-based code assistants three and a half years ago. Copilot was the state of the art back then and it was... what it was :). Code Whisperer was the other alternative, and it was scary bad at the time. You fought AI more than it helped you. Things slowly crept forward over time. Then, last summer, Sonnet 3.7 dropped, and something clicked. For the first time it felt like the model actually understood what I was trying to do across a whole codebase, not just the 40 lines of code on my screen. That was the lightbulb moment. Anthropic themselves called it the cleanest capability jump of that generation, and AI-assisted coding took off, with Claude Code dominating the field. We were early adopters, and never looked back. I thought it would take the frontier models another year or two to make the next meaningful jump. However, fast forward to today. Opus 4.7 is ASTONISHING, and Claude Code has become an absolute powerhouse. I've been working with Opus 4.7 and a Claude Code that's a Swiss Army knife of engineering, and the productivity speeds and near-zero error rates are literally flooring me. And Opus 4.7 is the neutered version. Mythos, the restricted model Anthropic is holding back for cyber safety reasons, is meaningfully more capable. Let that sit for a second. If you're still deciding what to use, a handy guide: Tier 1: Claude Code. No question. The gap is massive and widening. 46% most-loved (2x Cursor, 5x Copilot), $2.5B ARR, ~4% of all public GitHub commits. Most important - it works SO WELL, and it gets better every day. I think the race is theirs to lose, now and in the future. Tier 2: Cursor. I think the entire "Cursor is dead" parable is wrong. Different workflow, real traction, and still growing. Devs use it alongside Claude Code and I think it still has promise. Tier 3: Codex. 3M weekly actives, 60% of Cursor's usage. Strong at async background work. Loved by data scientists. Middle Ground: Copilot (incumbent), Gemini / Antigravity, JetBrains Junie. There's a "Copilot is useless" theme out there, which, comparatively, is sort of true. Copilot wins on ease of use, but that's it. These tools are fighting for relevance. Niche: Lovable, Replit, Bolt. This is the "not-for-engineers lane", which I happen to like, because it welcomes everyone to the software builder table. Legit purpose, different buyer. Falling off: Devin and Windsurf. Devin just never hit the promise: 14-of-20 failure rate in early evals, followed by price cuts from $500 to $20. Adoption's basically stalled, and I don't know any serious technology organization that is deploying Devin at scale. Windsurf: low adoption, and if you're looking for an AI-native IDE, you're choosing Cursor. One graveyard, two tombstones. TL;DR: if Claude Code isn't your default yet, you're leaving a generational productivity gain on the table. And if Mythos is directionally representative of what's next, we haven't seen anything yet.
-
Nick Ciubotariu shared thisAI's not replacing humans anytime soon. But man, humans have to step up their game in actually testing what they're yolo-codingNick Ciubotariu shared this
-
Nick Ciubotariu posted thisSome really uncomfortable truths coming: The way we "cost and measure" software development is basically toast. Story points are dead. Sprint velocity: dead. And if we're being honest, most of the Agile/SAFe/SDLC apparatus is basically meaningless (with "Waterfall" having mercifully died long ago). I was at a CTO conference this past week and this topic came up. The entire way we've estimated and costed software for the last 20 years assumed a roughly stable relationship between human effort and output. Engineer works X hours, produces Y output, and you can plan around that. AI completely broke that relationship. Now? A senior engineer with the right AI tooling can mass-produce a feature in two hours that would have been a two-week sprint six months ago. However, an architectural decision still takes the same amount of judgment time it always did. The unit of work has decoupled from the unit of output, and nobody's measurement system has caught up. Two-week sprints (or whatever timeframe) assume work arrives in human-sized chunks that take human-sized time. When AI can execute a "sprint's worth" of code in an afternoon, the sprint boundary becomes arbitrary overhead. SAFe's PI planning assumed you needed 8 to 12 weeks of runway because cross-team coordination was the bottleneck. AI crushes execution time so dramatically that you're spending (wasting) two days in a room estimating work that might take a tenth of what everyone just agreed to. Velocity tracking assumed consistent team capacity sprint over sprint. Now your velocity is a function of which tasks happened to be AI-compressible this sprint versus last. The variance is so high the metric tells you literally nothing. Definition of Done, acceptance criteria, the whole spec-first contract between PM and engineering: all that assumed extensive upfront specification. When the cost of building drops by 5 to 10x, it's faster to build three versions and pick the best one. The negotiate-then-build model inverts into build-then-evaluate. Here's how I'm now thinking about this: - The best we can do right now is complexity tiers, not points. Bucket work into trivial (AI does 90%+ autonomously), standard (AI-assisted, human-architected), and novel (genuinely new problem-solving where AI is marginal). - Measure time-to-capability over headcount x months. The question isn't "how many engineers for how long." It's "what's the smallest team that can deliver this capability by this date with acceptable quality risk." - Features delivered vs. committed roadmap is really the only accurate measurement we've got right now. - Be ready to to pivot to something else. This is the best we can do FOR NOW. AI can change this storyline next week. No one has answers, really, about how to predictably cost and measure in the fast-evolving age of AI. But we've got to stop pretending what we are using still works. The old model is gone. We need a new one. And right now, nobody has it. #ai #software #SDLC
-
Nick Ciubotariu shared thisThe era of predictive, agentic-based intelligence is coming. ShipStation is and will continue to lead and stay ahead of the curve here. Super proud of the team for their hard work in shipping this game-changing release for our customers! #ai #data #intelligence #auctaneNick Ciubotariu shared thisToday, we introduce ShipStation Intelligence — the data and AI layer powering smarter decisions across ShipStation. 🚀 It unifies analytics, automation, and performance insight into a single intelligence layer. Learn More ➡️ https://lnkd.in/gsdHBCnc Read the full announcement ➡️ https://lnkd.in/g3ij75qK
-
Nick Ciubotariu shared thisSo proud of my good friend Moshe Lieberman - incredibly excited to follow this story and wishing Steward lots of success!!Nick Ciubotariu shared thisInvestor onboarding in financial services is still far too slow. Onboarding complex investors - funds, trusts, family offices - is still largely manual. A small percentage of cases can consume the majority of compliance team capacity, while regulatory pressure keeps rising. That’s broken infrastructure. And exactly the problem Steward is solving. Excited that Motive Partners is leading Steward’s $5M funding round alongside Outward VC, Cooley and an outstanding group of fintech operators such as Shai Wininger (Co-Founder, Lemonade/Fiverr), Mark Ransford (Fintech angel investor), Tom Keiser (former COO of Zendesk and Carta), Remy Astié and Ulric Musset (Co-Founders, Vauban), Mushegh Tovmasyan (Founder, Zenus Bank), and Keith Grose (UK CEO, Coinbase). Founders Arik Oslerne and Moshe Lieberman know this problem firsthand. Arik saw where onboarding breaks down during his time as COO at Vauban, while Moshe helped build Lemonade and Fiverr from day one - companies that turned painful processes into seamless digital experiences. This isn’t another copilot. Steward owns the workflow end-to-end - embedding risk logic, handling complex ownership structures, and escalating intelligently when needed. Steward is already trusted with $100B+ in investor assets, using AI to automate complex AML/KYC workflows and enabling day-zero onboarding in 80% of cases. Exactly the kind of infrastructure financial services needs. Congrats to Arik, Moshe and the entire Steward team - excited to be on this journey together. Read the whole story on tech.eu and learn more about why we invested under the link in the comments. #Fintech #AI #Compliance #Venture #MotivePartners
-
Nick Ciubotariu posted thisEvery week, the frontier models' press machines run this now comical race against one another. Anthropic is the one company that's consistently delivering on what they're advertising. In software development, they have pulled so far ahead that I think it's soon going to be impossible for anyone to catch them. I'm a heavy consumer of Claude Code (and I complain about it a lot :) ), but nothing does what Claude Code does. Not even close. We massively accelerated Auctane by betting on it and scaffolding it to make it better, and same goes for my personal projects. I'm a dev, so design and PM skills are unnatural to me :). I'm now using Claude Cowork combined with Claude Code. And my God is it magical. Since eternity, the design-to-code pipeline moved in one direction. Designer hands off. PM translates. Engineer interprets. Context dies at every handoff. Claude Cowork, skills and plugins completely changed the game. - Design: Push a prototype to Figma and it lands as fully editable layers, not screenshots. The designer makes changes, uses the Design plugin, and pulls the winning design back into the codebase. Canva handles decks, stakeholder presentations, and marketing assets directly from Cowork. Start anywhere, go everywhere. - Product: the PM points Cowork at documents, folders, design artifacts, user research, Jira tickets, and Slack threads. Says "/write-spec." Gets back a fully structured PRD with user stories, requirements, and success metrics. Sub-agents review it from three angles before any human sees it: engineering feasibility, business alignment, and UX gaps. - Engineering: Cowork and Claude Code share a project folder. Claude Code reads the PRD, pulls design context directly from Figma MCP, and scaffolds the project to spec. Jira and Confluence connectors keep tickets and documentation in the loop. The engineer stays in the terminal. The context comes to them. Here's the really good part: this is a self-reinforcing flywheel. Engineer ships a build, designer captures it back into Figma, PM refines the spec, Claude Code implements the next iteration. How freaking cool is all that? What makes this work at scale is connectors and plugins. There are so many. 50+ connectors now (Jira, Confluence, Google Drive, Gmail, Slack, Figma, Canva, and more), all instantly available. You customize them for your stack, your terminology, your processes, and suddenly Claude works like it was built for your org. You can also build your own custom plugins for pretty much whatever you need. This works for new features and legacy code. In either case, something like this used to take tons of resources, funding, months, back and forth, delays, and now it's a fully automated process flow in a matter of days. The way software ships end to end has changed, forever. The companies that nail this loop first are going to ship and WIN at a pace that makes companies that haven't adopted AI in depth very uncomfortable, very fast. #ai #claude #cowork
-
Nick Ciubotariu posted thisAt Auctane, we used AI to re-write the entire Stamps.com front-end codebase: 30 years of legacy code, multiple frameworks, tons of tech debt, with four engineers in about 4 months. Goes fully live in a few days. The results: massive performance improvements, brand new codebase, feature acceleration, millions of lines of code reduced (~91%), modern architecture. Zero downtime, zero feature roadmap/customer impact. Here's how we got there: - We went all in on AI early and always met it where it was. AI isn't perfect, but it does some things incredibly well. Understanding complex code is one of them. We started with tasks AI was good at "at the time," then grew into feature development, code reviews, and design/architecture as the technology matured. - We learned from our mistakes and turned them into strengths. For instance, when AI hallucinated, we didn't just ask why. We investigated, built skills, guardrails, protective software, etc. Everything was an opportunity to get better. - We RELENTLESSLY focus on AI safety. Giving engineers AI is the equivalent of giving people nail guns. Speed and velocity go way up. So do emergency room visits. Thus, our operating mantra is: “Engineers own the code". No matter who or what writes it, humans are responsible for what ships to production. Period. There is always a human in the loop for perspective, structural thinking and quality control. - We eased engineers in, then quickly made AI use a requirement across the organization, not only in writing code, but in scaffolding the infrastructure, tools, skills and workflows that make AI effective and, in many cases, save it from itself. That last part is critical: we're not just deploying AI code editors for the sake of vanity metrics. Anyone can do that. ***We're scaling and building an enterprise-wide AI platform that stands up to production rigor.*** We train and learn. Relentlessly. Weekly AI office hours, Slack channels, AI champions, and we bring folks in from Anthropic, Amazon, Google, etc. to share best practices and learn from us as well. As a result: 100% of Auctane engineers use AI, a large part of our commits are AI-assisted, code quality is UP, and we've seen a 200%+ product velocity increase in the past year. And our engineers are better prepared, learning valuable new skills, and can focus more on engineering, product development and architecture rather than raw code output. This is what AI success in the enterprise looks like. In a world where there's so much talk about AI killing software, this proves the exact opposite. AI doesn't replace great software. AI is an incredible leverage multiplier for software that has strong orchestration, smart workflows, and a relentless focus on usability. The magic is, and will be, "better together". Last - a heartfelt "thank you" to an amazing engineering and product organization that started with "let's go, we can do this" and exceeds that bar every single day. (written by a human) #AI #Auctane
-
Nick Ciubotariu reacted on this🚀🚀🚀Nick Ciubotariu reacted on thisConsistency pays off! For the second year in a row, we're proud to announce WWEX Group has been named a Top 4 Freight Brokerage Firm by Transport Topics. This recognition is a reflection of our incredible teams, partners and customers who make it all possible #WWEXGroup #Freight #Logistics #SupplyChain #Top4
-
Nick Ciubotariu reacted on thisNick Ciubotariu reacted on thisWe're excited to announce that Episode 2 of The Technology Principal Podcast is now live on Spotify and YouTube! We're incredibly grateful to have welcomed Albert Ko, CEO of Auctane, for this episode and to have had the opportunity to learn from his perspective and leadership journey. In this conversation, Scott and Al explore what it takes to build durable leadership in today's evolving technology landscape - from overcoming challenges and paying it forward, to staying deeply focused on customer problems, operating with discipline, and understanding the enduring value of human judgement in the age of AI. This episode is a powerful example of what The Technology Principal is designed to create: honest, thoughtful conversations with leaders whose insights can genuinely shape how others think, lead, and innovate. A sincere thank you to Al for his time, candour, and the depth of insight he brought to this discussion. There are countless valuable takeaways in this episode, and we're looking forward to hearing what resonates most with you. Links are in the comments below. #Podcast #TechnologyLeadership #CEO #Innovation #AI #Leadership
-
Nick Ciubotariu reacted on thisNick Ciubotariu reacted on thisICYMI - the next episode of The Technology Principal Podcast goes live next week! This time, Scott Simons is joined by Albert Ko, CEO of Auctane - the global delivery experience company behind ShipStation, Stamps. com, and Metapack. In this episode, Scott and Al explore: - Overcoming (leadership) challenges and the importance of paying it forward - The important role human intelligence plays in the age of AI - What defines durable leadership in times of transformation For leaders navigating strategic change, digital transformation, and the increasing complexity of modern business, this is a conversation worth listening to. #Podcast #TechnologyLeadership #CEO #DigitalTransformation #AI
-
Nick Ciubotariu reacted on thisNick Ciubotariu reacted on thisWe’re pleased that the SEC has approved Nasdaq’s proposal to extend U.S. equity trading to 23 hours a day, five days a week. This approval reflects growing global demand for always-on markets that uphold the highest standards of operational excellence and investor protection. Our approach was shaped by close collaboration across the industry to ensure that any expansion of trading hours is implemented with appropriate safeguards and resilience, supporting orderly markets and effective price discovery. We look forward to continuing to work with regulators and market participants to expand access to U.S. equity markets.
-
Nick Ciubotariu reacted on thisNick Ciubotariu reacted on thisThe NASA - National Aeronautics and Space Administration Artemis II crew has touched down at Nasdaq MarketSite. Fresh off their historic mission, Commander Reid Wiseman, Pilot Victor Glover, and Mission Specialists Christina Koch and Jeremy Hansen joined CNBC Squawk Box to talk all things space — from Artemis II takeaways to what's next for humanity's return to the Moon.
-
Nick Ciubotariu reacted on thisNick Ciubotariu reacted on thisNothing That’s Cody’s entire answer to what he doesn’t like about ShipStation. In a world of dynamic carrier pricing, multi-carrier rate comparison is the difference between a profitable order and a break even one. Efficiency is table stakes, margin is the point. What’s the ROI on your shipping software? ShipStation #ecommerce #shipping #smallbusiness #gtm #saas
-
Nick Ciubotariu reacted on thisNick Ciubotariu reacted on thisPlease (please!) do not make me slog through your AI-generated slop. I know you are very impressed with the deliverable that Claude wrote for you. I'm sure it contains lots of good ideas. But it's exhausting and time-consuming to have to read it and decipher what you're actually trying to say between emdashes and "it's not X it's Y" cliches. If you gave Claude 3 bullet points to generate your document, just send me the three bullet points! We urgently need new social norms around this. In the meantime, please resist the urge to send slop. Thank you for listening, Eric
Experience
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Education
Volunteer Experience
-
Volunteer
Habitat for Humanity International
- Present 22 years 1 month
Poverty Alleviation
I happily donate my time to Habitat for Humanity to help build homes for those in need
Publications
-
Troubleshooting Common System Center 2012 SP1 Virtual Machine Manager Issues
MSDN/Channel 9
Best practices on how to identify and fix common WMI, WINRM, networking, host, storage and performance problems in System Center 2012 SP1. Troubleshooting tools, problem isolation tips, identification and location of various logs locations and guidance on making debug view logs helpful.
Other authors -
Patents
-
Alexa Comms - Voice-based Communications
Issued US 10,074,369 B2
See patentSystems, methods and devices for escalating voice-based interactions via speech-controlled devices. Speech recognition combined with Automated Speech Recognition (ASR) and Natural Language Processing (NLP) and Natural Language Understanding (NLU) techniques enable speech based user control of a computing device to perform tasks based on the user's spoken commands
-
Customer Impact Alarming Service
Filed US P42305
-
Streaming Authenticated Encryption
Filed US P39106
-
Alexa Comms - Speech-based Visual Indicator of Incoming Message #2
Filed US PM13656
-
Alexa Comms - Speech-based Visual Indicator of Incoming Message
Issued US 20180061403
-
Real-Time Latency Prediction for Customer Experience Optimization
Filed US PM11522
Projects
-
Visual Studio Online - Application Insights
The Application Insights project is bringing a decade of experience with monitoring and visualizing application health for the IT Pro to the developer who now has DevOps responsibilities. Offering this as a service, this enabled the DevOps with a low overhead, streamlined user experience to help them understand the performance and usage of their applications so they can deliver a constantly improving solution.
Other creators -
-
hirebuddy
-
Founder and Chief Technology Officer of hirebuddy.com
Other creators -
Honors & Awards
-
Microsoft High-Potential Employee Program (Top 1% employees)
Microsoft
-
Microsoft High-Potential Employee Program (Top 1% employees)
Microsoft
-
Microsoft High-Potential Employee Program (Top 1% employees)
Microsoft
Languages
-
Romanian
Native or bilingual proficiency
-
Spanish
Limited working proficiency
-
German
Limited working proficiency
Organizations
-
Arizona Technology Council
Council Chairman, Cloud Compute Committee
- PresentFounding Chairman of the Cloud Compute Committee for the Arizona Technology Council.
Recommendations received
32 people have recommended Nick
Join now to viewOther similar profiles
Explore more posts
-
William R.
If there’s any low level coders in my network please check out Troy’s GitHub and his project to build a new, faster shell! Some details on Troy: ### Windows Accessibility and TTS Contributions (Red Eagle Team) - Description: As part of the Red Eagle Team (a Georgia-backed initiative with Microsoft support in the mid-2000s), Mallory developed foundational code for universal Text-to-Speech (TTS) systems and cross-platform handicap accessibility features. This included compatibility layers for screen readers, app navigation, and voice input (e.g., influencing tools like Dragon NaturallySpeaking). - Impact on Windows: Microsoft adopted and evolved his logic starting with Windows Vista, integrating it into core OS features for impaired users. If you've used modern Windows TTS, clipboard integrations, or basic accessibility navigation, elements trace back to this work. - Modern Extensions: He's currently exploring IoT/smart home integrations for disabled veterans, building on this foundation. Troy’s Post “Okay guys think tank. I maybe over complicating this or may habe to do a bit of tweaking with some C code. Rubian has been going smoothly. The new boot and dynamic system discovery deamons work great. As standalone. The new stack needs tweaking for efficiency but works and is stable. Here's the stack im working on with the prototype test. /etc/rc.local ->calls boot.rb as bg process. ->boot.rb bootstraps rubian's system discovery and array logic. User login -> rubian3.rb (repl) -> chsh -s ./rubian3.rb USER Rubian boot.rb logic. ->starts C ruby and Jruby C ruby starts the bootstrap process while jruby warms up. -> Cruby passes workload to jruby when its ready to finish boot system. The deamon system is the server-side, where shells and env are the client. (See my UNIX socket prototype here) https://lnkd.in/gf6rWs46 C and j ruby can now natively talk to each other. I know that works. The issue im having is the new login shell isnt hooking into the deamon when it loads. https://lnkd.in/g5_2ZuPH A further explanation: “Yes. I've been building from the Shell working down to the kernel. The Shell side of things mimic Bash but in ruby expression. On the logic side of things, ruby treats everything as an object. What im doing is taking all the system information and building Arrays. Each directory and file then has an index that we can iterate through with .each_with_index, .each_with_object, etc. So instead of a static file tree, i have arrays of nested file paths and directories being pattern matched. Any ruby app will be ran and communicate to it. This allows for dynamic file utils systems and more. Each array index points to a file, directory, process, thread, interpreter, shell, hardware, everything that is a defined method object.”
-
Adriano Devillaine
Today we announced new capabilities for our Shop Direct and Buy for Me experiences on Amazon, enabling now feeds to facilitate external merchants to list reliably their product listings and reach millions of Amazon customers. Excited on this feature, that is a win/win experience for customers, merchants and brands. Excited to be part of such important project.
-
Marc Brooker
People often ask me about the differences in architecture between Amazon Dynamo (the 2007 SOSP paper), DynamoDB (the AWS serverless NoSQL database), and Aurora DSQL (the AWS serverless SQL databases). I captured the key points on durability, high availability, programming model, and trade-offs in my blog: https://lnkd.in/gkMG_tVD TL;DR: Despite the shared name, Dynamo and DynamoDB have little in common architecturally, and make very different trade-off decisions.
8 Comments -
Jay Jayapal
When James Gosling, the father of Java, says “AI isn’t the answer (yet),” it’s worth pausing. His commentary on Amazon’s recent challenges underscores a truth many technology executives know well: AI succeeds only when paired with strong fundamentals—architecture, governance, and human expertise. As we integrate AI deeper into enterprise systems, let’s remember that reliability and resilience aren’t by-products of intelligence—they’re engineered by people. #AI #GenAI #AIOps #ResponsibleAI #DigitalTransformation #EnterpriseAI #Leadership #TechnologyStrategy #Innovation #FutureOfAI
1 Comment -
Paul Kerrison
You could watch an entire episode of a sitcom, bake a medium-sized potato, or deeply contemplate the meaning of existence in 42 minutes. You certainly shouldn't be spending that long waiting for a CI/CD pipeline to finish. Slow pipelines are the absolute worst for killing momentum. It isn't just about the lost minutes, it’s the dreaded context switching. By the time the tests fail, you’ve completely forgotten what the code was supposed to do in the first place. However, fixing this doesn't always require massive infrastructure overhauls. Often, it’s about identifying bottlenecks, parallelising what can be parallelised, and swapping out sluggish tools for modern ones. It is about working smarter, not harder. This week’s tale from Nick Theodoulou dives deep into the "how," showing exactly how they used data, not guesswork, to reclaim over 100 developer hours a year. https://lnkd.in/eWYCNZA4
3 Comments
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content