David Long

Did you know there are no "roles" (SDE, SDET, TPM, …) when you get to L8 IC (Senior Principal) at Amazon? It’s one of Amazon’s peculiarities. [Side note: in general, Amazon SDEs tend to be "jack of all trades" where we pragmatically do whatever is necessary to make our organization successful at that given moment (within reason). In fact, I believe being well-rounded and able and willing to wear a number of hats is one of the things that makes me valuable to an employer. We don’t over-index on "that’s not part of my role doc." Other companies where I worked (eg Google) have very clear swimming lanes and people stick to them pretty religiously, eg. "oh no I don’t talk to customers, I’m an SDE and talking to customers is for TPMs." Whereas at Amazon I never once had a manager tell me, "what you did (although impactful, important and necessary) was not really your job." ] My little side note about this critical piece of amazon culture explains why as you get to L8 the concept of a "role" no longer really makes sense. The boundaries between roles get fuzzier and fuzzier as you go from Senior Engineer (L6) to Principal Engineer (L7), and they become unimportant when you get to L8. We’re all just “Senior Principals” - whether you came via the SDE path, the TPM path, or in my snowflake case, the SDET path. In fact I think most L8s are snowflakes in that their path to L8 was unique and the value they bring to the company is also unique. In phonetool (our amazon directory) you can have whatever unofficial title you want (also within reason...), so I chose to keep the "SDET" title to reflect my background and passion for testing, engineering and operational excellence, and developer tools. I also hope that by keeping the title, others in my (old?) role can see that SDETs at Amazon do not have a "ceiling" in their career. And lastly, it's a fun conversational piece ("you're an L8 SDET???") (The image here is "Peccy" - a beloved little icon that has represented Amazon's Peculiar Ways for years) [Edit: Strictly speaking there are a few L8 IC roles that do have a different job code, like PMT or Scientist - but I think we generally view all L8 ICs in the same lens]

929

41 Comments

How do you work through feeling overwhelmed with projects, both at work and at home? My project list is never ending. I have multiple major work areas, a house with a sheet of improvement projects that keeps growing, half a dozen trips planned into next year, and a wife and two daughters with lists of their own. It's easy to not know where to start or what to do next. And then before you know it you've burned an entire day playing sudoku. Many years ago I learned the art of Getting Things Done. I even wrote an article on it for USENIX Association's ;login: Magazine just as the world came screeching to a halt in early 2020, and talked about it on The Hedge Podcast (episode 88). There is a lot of structure to GTD, but I can boil it down to having a trusted list of all the next actions I need to do. One that is reviewed regularly and captures thoughts as I have them. All of it - work, home, family - gets treated the same way in one place. I've been out of practice and I've recently rekindled my system. This is one of the ways I'm addressing feeling adrift. I can always find something to do that will move me forwards, as well as give me that dopamine hit of crossing a task off my list. And if I'm avoiding one it's probably because I didn't break the next action down to a small enough chunk. The most important part of practicing GTD (and other organizational systems) is that it clears my head. The human brain isn't really great at maintaining lists. The only real method it has for doing this is to constantly go over the list, which keeps the pathways in the brain fresh. That takes mental capacity, whether you realize it or not, and it adds to your stress. Plus, half the time you don't even have the list in mind or complete when it needs to be. You think about the dead batteries in the flashlight when the power goes out at home, not when you're standing in Home Depot wondering what else you need to get.

22

1 Comment

Over the years I’ve seen a lot of things in Silicon Valley. One thing that hasn’t changed much? People making the same mistakes about equity compensation, taxes, and risk. That’s probably because the Bay Area has so many talented software engineers, who tend to have a “do-it-yourself” mentality. However, when dealing with complex matters where human behavior and specialized knowledge play crucial roles, a collaborative 'buddy system' almost always leads to better outcomes. 👥 If you'd like to work on any of the 5 items below and want to try out the 'buddy system', send me a DM. 🤝 #rsus #taxes #microsoft #meta #apple #google #Nvidia #amazon #tesla #magnificent7 

4

5 Comments

Tesla’s shift to robotaxis marks a strategic pivot to become a service-oriented company, mastering the entire spectrum of cars, tech, and energy. This move transcends even their upcoming $25K cars, as meeting transportation needs could reduce the demand for private car ownership. I believe in practice Tesla could double-down on the long-term vision. #tesla #robotaxi https://lnkd.in/eTbS56c5

3

3 Comments

Judge a company and its leadership not by how they treat people when times are good. Judge them by how they treat people when times are hard. Ive seen many, many examples of “we need to institute a performance culture” (compared to what, the non performance culture you, Mr CEO, were responsible for prior?). I’ve seen “we need to trim the fat” (who you, Mr CEO, allowed to be hired). Work from anywhere, as long as you get results (until we want to reduce headcount without paying severance). “We’re a family where you can bring your whole self to work” (hint: work isn’t a family, and I’m pretty sure even the HR person at orientation now might say “leave your emotional support alligator at home, please”). Imaging putting a family member on a pip (sorry grandpa, unless you meet these 3 OKR metrics next month, hit the bricks buddy). It’s easy to spike the football when things are going great. The fact Nvidia didn’t copycat most of the rest of tech by going into “manage by fear” mode in 2022 after the stock went down by 70% is why Jensen is one of the most admired and respected leaders in business.

115

2 Comments

AI drivers now outperform humans. Drastically! . . 22 million miles, zero drivers, 84% fewer crashes. AI just crossed the Rubicon: Waymo's safety data reshapes autonomous driving forever. We probably just crossed a historic threshold in autonomous driving safety. Buckle up, because this changes everything. Waymo's latest safety data isn't just impressive—it's a game-changer. Here's why it matters: 84% fewer airbag deployments 73% fewer injury crashes 48% fewer police-reported crashes And they're not just cruising empty streets. We're talking 22 million miles of real-world, hands-off-the-wheel driving. But here's the kicker: Waymo's now serving up 100,000 robotaxi rides weekly. That's not just a number—it's a seismic shift in urban mobility. Think about it. We're witnessing the birth of a new era where: a) Cities evolve around AI-driven transport b) The concept of car ownership gets a complete overhaul c) Insurance models flip on their head The ripple effects? Massive. We're looking at: a) Reimagined urban planning b) A potential boom in mobility-related tech c) A complete rethink of traffic laws and road design But let's zoom out. This isn't just about safer roads or convenient rides. It's about unlocking human potential. Imagine a world where commute time becomes productive time. Where accessibility isn't a privilege, but a right. The questions that keep me up at night: a) How will this reshape our cities in the next decade? b) What new industries will emerge from this revolution? c) How do we ensure this tech benefits everyone, not just the privileged few? What's your boldest prediction for the autonomous driving revolution? Let's think big—the future is autonomous, and it's arriving faster than we thought. #AutonomousDriving #AIInnovation #FutureOfMobility #urbanmobility #Waymo #Selfdriving #AI #robotaxi -------------------------------------------------------------------- ♻ Gained an edge from these AI insights? REPOST and share the boldest AI-driven business moves with your community.

2

The "studies" that show San Francisco in a bad light are like comparing gradeschoolers to college kids. Nampa, ID is the best? They have 110K people and a budget of $279M. It's basically a suburb of Boise. SF has 808K people - the 4th largest city in the US - with vastly more complex issues like public transit, homelessness (you can't freeze to death here, unlike Idaho). SF is also a county. It's overall budget is $14B. Comparing metrics between these two is disingenuous at best. We have to start calling this kind of reporting out. Facts matter. None of the cities in this lists top 20 are major metros. SF has government problems, but this kind of hyperbole is just slamming on SF. I wonder why? https://lnkd.in/e8qnGeg7

6

3 Comments

World class writing as engineer is the top skill that differentiate and demonstrate the value.

7

At Instrumental Inc. we are typically very secretive about the clients we work with. During onboarding 3 years ago, I knew I made the right decision to join, especially after seeing the list of major logos we support. Today we are doing something a bit out of character and publicizing our partnership with @Meta to assist in the quality control of their new hardware. I personally believe this is the sort of edge hardware companies need to outperform competitors in a rapidly evolving market. Don't just take my word for it though, check out the video! Really exciting chapter in tech and at both companies. #AR #Engineering #Innovation

14

In the early days of the internet, the W3C and few other organizations helped build the first Internet Standards so that the Web could foster it's own innovation & growth. Fast forward 30 years, I cannot believe we needed a law to get browsers and websites to agree on protecting your privacy by default... The internet should really be decentralized, and customer centrist by design. How did we ended up here?

11

🚀 𝗠𝗶𝗻𝗶 𝗔𝗜 𝗠𝗼𝗱𝗲𝗹𝘀: 𝗛𝗼𝘄 𝗱𝗼 𝗺𝗶𝗻𝗶 𝗺𝗼𝗱𝗲𝗹𝘀 𝗮𝗰𝗵𝗶𝗲𝘃𝗲 𝗢𝗽𝗲𝗻𝗔𝗜-𝗹𝗲𝘃𝗲𝗹 𝗿𝗲𝘀𝘂𝗹𝘁𝘀? 🚀 As the AI industry continues to grow, a new wave of mini AI models is challenging the notion that bigger is always better. Recent research from the Allen Institute for AI (Ai2) has shown that smaller models like Molmo can match or even outperform larger, proprietary models like OpenAI’s GPT-4o, but with 1,000 times less data. 🌟 𝗗𝗮𝘁𝗮 𝗤𝘂𝗮𝗹𝗶𝘁𝘆 𝗢𝘃𝗲𝗿 𝗤𝘂𝗮𝗻𝘁𝗶𝘁𝘆 - Molmo Models: With sizes ranging from 1 billion to 72 billion parameters, these models achieve competitive performance by focusing on high-quality, well-curated data instead of vast amounts of raw information. For example, they trained using 700,000 images and 1.3 million captions, a stark contrast to the billions of examples fed to GPT-4o. - This efficient approach means lower training costs and reduced energy consumption, making high-performance AI more accessible to researchers and startups. 🎯 𝗜𝗺𝗽𝗿𝗲𝘀𝘀𝗶𝘃𝗲 𝗕𝗲𝗻𝗰𝗵𝗺𝗮𝗿𝗸 𝗣𝗲𝗿𝗳𝗼𝗿𝗺𝗮𝗻𝗰𝗲 - Despite being 10 times smaller, Molmo’s 72B model matches the capabilities of OpenAI’s GPT-4o, Anthropic’s Claude 3.5 Sonnet, and Google’s Gemini 1.5 Pro across 11 academic benchmarks, including reasoning and language understanding. - This proves that model size doesn’t always equate to better performance, especially when smaller models are trained with targeted, high-quality datasets. 📈 𝗖𝗼𝘀𝘁-𝗘𝗳𝗳𝗲𝗰𝘁𝗶𝘃𝗲𝗻𝗲𝘀𝘀 𝗮𝗻𝗱 𝗔𝗰𝗰𝗲𝘀𝘀𝗶𝗯𝗶𝗹𝗶𝘁𝘆 - Training these mini models uses significantly fewer resources—up to 75% lower training costs compared to traditional models, thanks to their smaller data requirements. This means businesses can deploy robust AI capabilities without the need for massive infrastructure investments. - The shift toward open-source mini models is also democratizing AI, providing Apache 2.0 licenses that allow developers to adapt and use these models in commercial applications, free from the licensing restrictions of larger models. 🌍 𝗪𝗵𝘆 𝗧𝗵𝗶𝘀 𝗠𝗮𝘁𝘁𝗲𝗿𝘀 The emergence of these compact, efficient AI models is a game-changer for the industry. They offer a path to sustainable AI development without sacrificing performance, opening doors for smaller teams and innovative startups to leverage advanced AI capabilities. As AI continues to evolve, focusing on data quality rather than just data quantity may redefine what’s possible in this field. #AI #OpenAI #Molmo #MachineLearning #DataEfficiency #TechInnovatio How do you think smaller AI models will shape the future of AI? Let’s discuss!

2

🚀 How to Approach a System Design Interview: Amazon Prime Video Example When tackling a system design problem, especially for a complex platform like Amazon Prime Video, it's crucial to approach it strategically. Here's how I break it down: Get clarity on the requirements, both functional and non-functional. Focus on the user flows, define the scale, detail out the components, improve scalability, and wrap up confidently. Pro Tip for Interviews: Don’t rush into drawing diagrams! Ask clarifying questions, understand the problem fully, and then build the system piece by piece. Communicate your thought process clearly, and focus on scalability, reliability, and cost-efficiency. Remember, the trick is to think aloud—explain why you're making certain decisions. The interviewer wants to understand your reasoning as much as the final solution! Good luck to all those preparing for system design interviews! You've got this!💪 Want to master system design interviews and land those $500k+ TPM roles? Make sure you're fully prepared with the Kracd.com interview prep course. Sign up today and transform your career in just 30 days. #SystemDesign #AmazonPrimeVideo #TechInterview #InterviewTips #Scalability #CDN #TechCareers #SoftwareEngineering #Microservices #TechLeadership

9

A few thoughts after visiting the Hadrian library - Hadrian library was estimated to have housed 17000 books. This would make it one of largest in the ancient world. In comparison, the library of Alexandria was estimate to have 200,000 to 700000; Google Books 40 million. - Everyone of us have access to that much information in our pockets. 17000 books would fit on a 32 Gb Kindle Paperwhite. The internet is estimated to have about 150 ZB of information, more than 5 million times. - These cultural treasures/institutions, ancient and modern, are worth preserving. https://lnkd.in/g6vFq-tW - Consider donating (time or money) to organizations such as https://www.wikipedia.org/ and https://archive.org/

8

LLMs are not the Holy Grail of software coding. They are great for boiler plate code and getting quick answers to questions on documentation, but fall flat after that. Here's why... 1) You need to understand how to code to provide the right questions or prompts to an LLM 2) LLMs don't take your entire codebase into context so you end up with duplicated, sloppy code. Yes, you can mention or highlight relevant parts of your code to help the LLM create a better answer, but at that point you're kind of helping the LLM cheat 😀 What do you think?

3

1 Comment