Chris Smith

Agree that MS/GH have moved extremely quickly to incorporate LLMs into their products -- but I think that's the key word, incorporate. This isn't disruption in the traditional sense. Copilot Workspace largely does things that GH already did in some form -- you could make plans in GH Issues, you could run apps or pieces of apps in GH Codespaces, etc -- but uses LLMs to enforce hygiene and walk devs through the ideal version of the process they were already more or less doing. Combining GH's existing tools with a super smart English speaking robot is indeed a recipe for magic. That's more or less what MS has done with Copilot too -- this is a case where the corporate mandate to "figure out how to use AI in your product" actually kind of worked, because it wasn't asking anybody to really rethink and reimagine their product or workflow. The speed at which they've done all this does show that startups should focus on problems or approaches that the big guys can't or won't -- if there's one weakness to GH Copilot, it's that they can't imagine a world where the developer isn't fully in control, both because that's GH's core ethos (developers are in control) and because they sell Copilot by the seat.

3

*Bracing myself for the response I might get for asking this* Software Engineers: Would you ever consider working on an EQUITY ONLY basis (ie no salary, just owning a portion of the business, the value of which would increase as it progresses and scales)? Assuming: - The product is really interesting and so is the work - The product is in the "tech for good" space and would be a really positive contribution to parts of society - Seed funding is confirmed to be coming with several backers keen & they will able to offer a salary in the near future - The product is super early stage, you would be writing the software from the ground up, but the concept, design and business model once funding is received are all solid I know that there will be many who would never consider this under any circumstances and I totally get why, but just curious to hear if any of you would if it was for a product you believed in enough? And would you look at it on a full-time basis or a part-time basis? It's something I got asked about and so I thought I would throw it out to Linkedin to see what you all think rather than assuming that I know the answer on your behalf. #startup #founder #softwareengineer

40

320 Comments

Honestly, being a good engineer isn't that hard. It is mostly taking that extra bit of time to weave quality into your work and not be lazy. It only takes 5 minutes to record a decent test plan of your code changes and attach it to your pull request to clearly show what your commit does. It only takes 5 minutes to answer your junior teammate's question and make them more independent instead of ignoring them. It only takes 5 minutes to create a set of meeting notes and set the agenda ahead of time to prevent the meeting from being a meandering mess. This isn't even going the extra mile, it's going the extra foot. You'll be surprised at how big a difference small optimizations like these can make. However, if you want to go that extra mile and become a great engineer instead of just a good one, join us for our upcoming educational session with engineering prodigy and Senior Pinterest engineer Jordan Cutler. In this session, Jordan will break down 21 quick tips for you to become a way better engineer. The event is open to all, but we are limited to just 500 spots due to Zoom. Claim yours here: https://lnkd.in/giFPg8ge #techcareergrowth #softwareengineering #growthtips

283

15 Comments

 You’d be surprised how many tech decisions are actually… business decisions So you have this company and C-level executives. I noticed some stuff, and I’m not the only one. Decisions such as: ⇒ Let’s use React! ⇒ Let’s switch the architecture towards microservices ⇒ Let’s deploy our app using Kubernetes You’d be tempted to think that all these decisions are of a technical nature… right? You’d think that the CTO is the only person who should make these decisions… right? WRONG! Most of the decisions mentioned above are masqueraded as technical. In real life, these are actual business decisions through and through. Why do I label them as business decisions? Because all three will impact the team, not just the technology you use. This is how: ⇒ Do I have the time to train the team? ⇒ Can I actually invest in training the team?  ⇒ How big of a community does the proposed technology have? ⇒ How will this impact the company in terms of support and ongoing maintenance? ⇒ How easy it is to maintain the new system? Will the costs of KTLO activities go up or down?  ⇒ How easily can I source the talent and/or attract them?  ⇒ How much will I pay for the support of those specialists on the market? For example, when you pivot towards Kubernetes, everything is more complicated. You’ll need specialized help, and it will cost you a lot more than your current solution, for which you already have available specialists. For a lower cost. The impact of the questions listed above is so major that, from my perspective, they are definitely business decisions, not technical. The CTO needs to be involved in the business. He should explain the impact of all the points listed above, and then follow-up questions will ensue. The debate starts, and you’re closer to the best solution for the company. That’s all, folks.

2

Using “average” latency is very misleading when monitoring any distributed system. Percentiles give you much more information to work with. I wrote a quick article to explain the difference this can make. 📝 https://lnkd.in/enaZpN25 #monitoring #observability

20

This video explains why I think Open Source projects, and the The Apache Software Foundation in particular are so powerful https://lnkd.in/eFZ3aXMY

52

Being able to effectively read software and system logs is a huge skill. Quickly grep'ing, correlating log entries, filtering out noise, and navigating multiple log sources is something I can attribute countless debugging and troubleshooting success to. Even just having the mindfulness and willingness to dig into logs is already a step in the right direction. Wise words from Davanum Srinivas 👏 #softwaredevelopment #softwareengineering #sre #sitereliabilityengineering

50

2 Comments

Adaptive In-Context Learning (ICL) meets Vespa. I wrote a blog post series on using Vespa's retrieval capabilities to retrieve context-sensitive examples for ICL. ICL, as I see it, democratizes applied machine learning, making it accessible for more organizations; all you need are good examples. https://lnkd.in/dijuSWBf

42

1 Comment

I’ve had some luck creating a #documentation pipeline using Doxygen and Moxygen to go source comments > XML > markdown. Moxygen needed a little tweaking to get the output I wanted, but it’s looking pretty good! #programming #cpp #devrel #docker

27

🔁 Balancing loops in software engineering are tricky. Push your team too hard, you get results, but only short-term. Relieving pressure? Short-lived contentment. There's a sweet spot in between. #SoftwareEngineering #SystemsThinking

1

🔁 Balancing loops in software engineering are tricky. Push your team too hard, you get results, but only short-term. Relieving pressure? Short-lived contentment. There's a sweet spot in between. #SoftwareEngineering #SystemsThinking

2

Arg! After thousands of lines of code cannibalizing the LLVM support library I’ve run into a fundamental design flaw in my approach. Too technical to talk about. Let’s just say I’ve created a set of cyclic dependencies where a depends on b which depends on c which depends on a. It’s because I made a fundamental change to how the library works (which makes it much easier to use when it has to work on different platforms like MacOS and iPads, and is the main reason I went down this path rather than trying to use the library as is) while trying to stick very close to the original structure and design. Not gonna fly. So either I throw out all that work and give up the ease of use - not happening - or I step back and rethink the big picture design elements. And despite the frustration I am loving digging into this library. Usually. By the way this is pretty much what programming is all about - running into obstacles and figuring out how to overcome them. We call it Software Architecture, but it’s not at all like building cathedrals.

6

5 Comments

Truth. This blog articulates the shortcomings of most tiered storage solutions in delivering predictable performance at scale. Without some rebuild that harnesses cloud primitives like AWS S3 and GCP GCS, mere incremental adjustments to existing products won't adequately serve users. From a product manager's perspective, sometimes incremental adjustments isn't a path to sustainable product adoption and success.

9

Don't waste your time with TDD. Writing tests-firsts is only necessary when... ... you want your coworkers to be able to refactor what you wrote this morning. The title is clickbait, but I appreciate your curiosity. Thank you for sticking along! I find that TDD as a whole is tightly connected with other Extreme Programming principles and disciplines. The outside-in-flow is going through the entire delivery cycle: 1. We are wrong about most requirements 2. So we want to deploy early and frequently 3. With features that were built just-enough and just-in-time 4. Expressed with written and automated tests 5. That match our understanding of desired behavior 6. So that anyone on the team can change it 7. When we figure out what we were wrong about TDD comes in between step 3 and 6. But you see... none of that makes sense if your engineers only ever write new code 95% of the time Rather than iterating over entire wholes with the team. #tdd #extremeprogramming 🚫 It's you vs. the algorithm—take control of what you see and join thousands of engineering leaders 🔔 Hit the 'view my blog' button under my name to subscribe 🔔 ♻️ Reposts feel like hugs

34

53 Comments

In a system design interview, you really only need to understand a handful of key technologies in order to do well. We've been working on creating a set of "to-the-point" breakdowns on each of these core technologies. We started with Stefan's great breakdown of Redis: https://lnkd.in/g3CQtDD7 And now I've just published a breakdown of the popular streaming platform, Kafka: https://lnkd.in/g6f3AjGy What do you want to see us break down next? Elasticsearch? S3? Something else?

42

6 Comments