Stuart Peschka

Does anyone else remember sitting in class thinking, "Am I ever going to actually use this?" 🤦‍♂️ I still ask myself this weekly when reading blog posts, listening to podcasts, catching up on massive Slack threads, or diving into documentation and feature listings. I don't always use the info, but it feels great when a memory from six months ago suddenly becomes relevant. I always think of that original Jurassic Park Unix scene. This often happens when discussing DevOps and Serverless, where boundaries blur already. Trying to keep all of the AWS documentation in my head while discussing the design of new features and optimizations feels like a superhuman task. What are some things you wish you had remembered when solving an issue or problem? For me, we designed a Lambda function to fail purposefully instead of attempting to recover for efficiency. Unfortunately, the Lambda control plane self-throttles a function once a failure threshold is reached... slowing our processing to a crawl and forcing some quick code changes. 😅 #AWS #lambda #serverless #learninginpublic #hugops

13

1 Comment

My #1 management rule that I’ve seen even very senior people get wrong - people’s actions will mostly mirror how you treat them: If you treat devs like children with blinders on they will act like children. If you hide costs from devs they won’t care what things cost. If you don’t have devs on the hook for when things break they won’t care so much about when they break. If you treat office workers as untrustworthy they will act untrustworthy. If you force RTO to keep an eye on people they will waste more time in the office. The tighter you squeeze your fist the more falls through. Thats not to say there aren’t exceptions to this but it’s true in the general case.

57

10 Comments

Scaling a team is often necessary as demand grows, but it’s essential to ensure it’s done effectively. Before adding headcount, consider if scaling is the best approach—sometimes addressing process inefficiencies or architectural issues can yield better results. If scaling is the answer, keep these factors in mind: * Define clear goals for each team. * Ensure accountability and support. * Minimize dependencies and avoid inefficiencies. * Scaling impacts people, process, and technology. Successful team growth requires careful planning to maintain efficiency and productivity. https://lnkd.in/eTkqykmG

15

4 Comments

This is a great video about how our personal ideology impacts the way we process data and information. For me, it shows the importance of fighting against populism and "framing" in the media and teaching our children in critical thinking. https://lnkd.in/eXQdB4fR

2

Leading a remote engineering team of 28 has taught me one thing—connection isn’t optional. It’s not just about keeping operations running; it’s about intentionally creating moments of connection and trust. As this article points out, remote leadership goes beyond virtual meetings and structured check-ins. Remote work doesn’t mean distant work—it’s just different. #RemoteWork #EngineeringLeadership #RemoteTeamManagement #Startups #WorkLifeBalance #TeamVibes

25

So true! In my opinion, every effort, whether it's new features or services, should include the creation of proper data and analytics to measure its effectiveness and adoption. The necessity of a feedback loop through telemetry and analytics cannot be overstated. It turns data into actionable insights, enabling iteration and refinement of the offerings. Without this, we risk repeating the same errors or just offering the wrong solutions. understanding the customers and having a constant feedback look is a must!

14

Many businesses claim to be tech companies—but are they really? 💡 In my latest Substack post, I delve into: The true definition of a tech company. Heuristics to determine if your organization fits the bill: Do you use data to drive decisions? Are you constantly automating processes to improve operations? Are technologists making strategic decisions at the highest levels? Drawing from my experiences at Carvana, we explore what sets genuine tech companies apart from those that simply use technology. #techindustry #innovation #leadership #datadriven #automation #techleadership #carvana

3

tldr: I'm hiring! As many of you know, I recently joined Lattice as their SVP of Engineering and I'm excited to announce that I'm hiring for all Engineering levels and more broadly, Lattice has open roles across EPD. If you're wondering, why Lattice? The last few years have seen a step change in how we think about the meaning of work and the employer <> employee relationship - add on the fuel of AI taking our jobs and you could be forgiven for thinking the future of work looks bleak for employees. It doesn't need to be that way - here at Lattice we're building the next generation of talent tools to give employers insights and employees purpose because people matter at work. Among other things, we're using generative AI to securely analyze enormous volumes of data to unlock career insights faster and investing in building equitable feedback, compensation and employer analytics tools - not just for one company but for all companies. If you're interested in making a difference to the future of work then let's chat - all roles are remote eligible and based in the US. We have physical offices in San Francisco and NYC if you prefer the in-office environment.

324

48 Comments

Let's 🎉 celebrate 🎉 the fact that AFLAC users have the ability to use passkeys to speed their logins, improving the user experience, and reducing their customer service costs. But I can't overlook this quote: "Passwords are still in place for users who are not ready to move to passkey, or as an alternate path if needed." In other words, users are still subject to the risks of shared secrets. The security needle hasn't been moved in a material way and phishing resistance hasn't been achieved. Progress is good, but let's not oversell the benefits of passkeys in deployments when we haven't removed the risk of shared secrets or weak account recovery mechanisms. What do you think?

32

10 Comments

"We thought we'd solved scheduling in Outlook when we invented 'free/busy' time... Fast forward two decades and now we have systems where one side structures a calendar, and the other side just picks an open time off a website. It’s not unlike what we did with free/busy. I don’t know about you, but I’m 100% on meetings being rescheduled when I pick the time off someone else’s calendar without their input." Automation is a gigantic PITA. If you think it’s not, you’re wrong. Automation is never straightforward. It's always handling exceptions—and almost everything is an exception. It's about abstracting way complexity - and every abstraction leaks. This is a good article. Give it a read: https://lnkd.in/gaRKwR6Z.

2

Some takeaways for a Jeff Hinton interview with timeless insights. My POV as an app builder is that models are now being improved for better UX (faster response/multi-modal) and reasoning capability and cost continues to improve. The technology headroom remains strong, and the integration of LLM into all software experience will continue and expansion of the reach (invasion) of software + intelligence will also continue. - Next word prediction is not merely parroting of training data. To do next word prediction well, LLM needs and already exhibits strong reasoning behavior. This would continue to scale / improve with more data / training. - LLM is a compression of human knowledge. The compression mechanism is identifying similar concepts across domains and reduces duplications among them. This can also be a source of "creativity" as LLM "sees" correlations of seemingly unrelated ideas through this compression. LLM can be more creative as it sees similiarity / correlations across all domains beyond (human) convetional wisdom. - LLM can exceed the quality of its training data. Jeff gave an example of an experiment where a model trained on 50% intentionally incorrectly labeled data, can still perform at 90+% accuracy after training. - Multi-modal models can improve the overall world understanding and lead to better overall reasoning. It does not even need to generate multi-modal output. - Not ML specific, but an interesting life insights. Bad intuition comes from believing (or wanting to believe) everything you are told. It diminishes the ability to make a clear choice that would lead to differentiated success (or failure).

4

Don't forget to budget for HiPPo removal in your 2025 budget! The #1 Reason strategy and initiatives fail? Alignment. One of the biggest reasons for the lack of alignment? Hippo's (highest paid person's opinion) or loudest voice or silo thinking etc ... dysfunction is an alignment killer. Business Outcomes are the secret Hippo (and dysfunction) Removal service. Business Outcomes Maps for an initiative declare "the why and the way" we will get to the True North for the initiative. Every Business Outcome and Investment area is prioritized. The Hippo's can roar all they want but the Business Outcomes Map is the committed and AGREED path we all signed up for. That Initiative Map and the agreed Business Outcomes gives even the quietest (smart) mouse in the room the RIGHT and CONFIDENCE to stand up for what the team has agreed. So in 2025, invest some of that budget for our HiPPo removal service. Your team, solution providers and board will be glad you did! And ... let more of the smart mice get on with what we're good at.

5

It's the first time I've seen that ChatGPT is making a typo in the generated code. This example is the 3rd version of the code I asked to generate, and only in the third one did it introduce a typo (outputWiorkbook instead of outputWorkbook) So, it learned how to make human errors.      static void ProcessExcelFiles(string inputPath, string outputPath)     {       using var inputWorkbook = new XLWorkbook(inputPath);       using var outputWorkbook = new XLWorkbook(outputPath);       var inputData = inputWorkbook.Worksheet(1);       var outputData = outputWiorkbook.Worksheet(1);

12

3 Comments

Software engineering is not the same thing as just coding, why don't people understand that?? Most of software engineering is actually thinking about and understanding the ecosystem/requirements. Once you understand that, writing the actual code is actually usually quite easy.. easy enough that it can be outsourced to AI maybe.

19

3 Comments

Reading through a discussion of when to automate tests (or not) on Ministry of Testing slack, I was reminded of this wonderful resource, Dorothy Graham and Seretta Gamba's Test Automation Patterns wiki. Check it out! https://lnkd.in/e9buqinn #TestAutomation

18

1 Comment

Early on the organic proliferation of synonyms used to describe or even define new software functionality seems innocent enough, but certainly has medium term risks. Naming is the origin of all things and there is an argument synonyms dilute the experience being built and make the development process less efficient.

1

There's an oft-cited stat from Pendo that 80% of software features are rarely, if ever, used. While it's impossible to know the accuracy of that, it's not hard to believe. I've been part of multiple feature factories, where emphasis was put on new features to 'claim' rather than taking time to ensure we've actually solved the desired problem before moving onto the next. If you're a Product Manager/Leader in the Indy area, join us on August 14 for a conversation around best practices for measuring product success. We'll have food, drinks, and hopefully some great learnings! (Registration link in the comments)

21

3 Comments