This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
Greetings from New Zealand, Kia Ora - excited to hear from our speakers and experts - Day 2; bring it on :kiwifruit:
Buenos días from an overcast Sierra Norte de Madrid. Looking forward to day 2 of the conference!
Reminder: Get yourself in front of your browser for the start of Day 2 of the Summit in 15 minutes at 10:00am BST! https://devopsenterprise.slack.com/files/UATE4LJ94/F06V49GSELX/timer.png
My favourite picture of Dev and Ops - (I lie, they're actually called Alfie and Bertie...) - hello from Leyland, North West England
My favourite picture of Dev and Ops - (I lie, they're actually called Alfie and Bertie...) - hello from Leyland, North West England
I hear John Allspaw is working on a new talk based on this “below the line” research finding 😉
A very early good morning from Portland, Oregon. And yes, the ITREV team is running this show live, and we would love to hear what you think of our new format. Cheers to a great Day 2!
Good morning from Indiana! 2am for you 😂 Excited for Day 2! 😊
I am really looking forward to Stephen Fishman and Matthew McLarty’s talk on optionality at the end of the day. I am not familiar with their work so I don’t know whether by optionality they mean a custom pick & mix candy bag of business features on call through API’s, or whether their connotation of optionality is about postponing the moment your options become decisions (to extend the candy metaphor if I may: to decide what type of candy bar you want the moment you feel hungry instead of many days before). I absolutely loved the graphic business novel about applying the concept of ‘Real Options’ in project risk management. The title of the novel is ‘Commitment’ (2013) and it is written by Olav Maassen and Chris Matts. See: https://commitment-thebook.com/. That book was more a traditional project management setting, I guess, and examples of applying optionality in a scaled agile setting are ‘Set-Based Design’ and ‘Thinking in Bets’ (Annie Duke). #discussion
Reminder: Get yourself in front of your browser because Day 2 of the Summit is starting now!
remember to take also a good Italian espresso and enjoy ☕ the day 2! 🎉
Single-track awesomeness creating a great shared experience
2*Emojis here to show interest in watch parties*
Maybe promote watchparties a bit more on the events page. I wasn’t aware of them…
🌅 Starting us off today is @willem.vanlammeren, Technical Lead Industrial IT, Solvay and @david856, Manager, Analytics for Industry; Author, The IT/OT Insider. They will present From Mass Manufacturing to Mass Innovation and tell the story of one of the most significant challenges in the manufacturing industry.
Our AV engineer @kalle.makela setting up the watch party in Tampere Finland! 💛
One of the most common things I heard when working with people in crtitical infrastructure was, “But it’s OT.” (meaning, improvement is impossible and it’s hopeless.) Which is why I think this topic from @willem.vanlammeren @david856 is so important.
I am still amazed by the under-representation of industrial engineering in IT and DevOps. Deming, Toyota Kata are all staples of IE.
I've long noticed, though, that there seems to me a larger-than-expected number of mechanical engineers among agile practitioners (I'm a mechanical engineer)
industrial engineer here :man-raising-hand:
2Interestingly, the people at the forefront of each Technology-led Revolution, since 1771, are Engineers. Whether mechanical engineers, industrial engineers or software engineers.
It's because good mechanical or electrical engineers understand the scientific method. Agile is based on the scientific method.
My systems engineering lecturer always said - A good engineer is a lazy engineer - they wil spend weeks to find a way to save 10 mins on any tedious process.
According to Larry Wall, the original author of the Perl programming language, there are three great virtues of a programmer; Laziness, Impatience and Hubris
i can imagine some of the conversations inside basf… “we need to redeploy our containers that manage our containers.” “we’re upgrading our pipelines that handle the pipelines.”
Good point: NotPetya virus prevented Maersk from being able to unload their physical containers from ships
I remember having some confusing discussions about "products" at a bank where financial products are often quite different to a digital SW product even though they heavily use a lot of tech
When I worked at BP I could use lots of nice analogies about Flow of work being like physical pipelines. Management's role should be a "pig" (https://en.wikipedia.org/wiki/Pigging) removing the crud in the pipeline (or even worse kinks in the pipeline) to increase flow. I've had to find a load of new analogies for Atom Bank 😁 .
The real world is atoms: clean water, manufactured goods., etc
…and they are a huge part of the economy, creating lots of good jobs
It’s not “Fail Fast” it’s “Learn Faster” Failure is just what some of us learn best from.
SCADA, PLC, sensors, networks (DCS), enable controlling entire plant from one control room. Look at all these happy people in this marketing slide — but it actually runs in Excel. 😆
First graph is good! But what is happening in the second graph? Manufacturing growth seems to have stalled.
(I love these graphs from FRED site: Fed Reserve of St Louis). Kudos to @willem.vanlammeren @david856! PS: I got a tutorial on this from one of Dr. Ethan Mollick’s old classmates, Dr. Daniel Rock, who spoke at GeneCon last year on AI impact on labor markets (the GPT for GPT paper that he did with OpenAI researchers) https://fred.stlouisfed.org/
“one third through pilot+ phase, management loses attention due to another shiny object”
Question is what is failure. It is very hard for people in manufacturing to accept to stop a project when investments were made (sunk cost fallacy). For me, stopping in time and learning from that experience can be extremely valuable
Your experiences in manufacturing happen across many industries. Just an example - Suncorp in Australia (financial services) went on with their core banking transformation for years despite not going anywhere ... sunk cost fallacy in action.
Amy Edmondson has been researching failure (and learning from it) for years: https://hbr.org/2011/04/strategies-for-learning-from-failure Her latest book is: 'Right Kind of Wrong: The Science of Failing Well'.
@joachimsammer - Hadn't seen your reply before, but thanks for sharing this HBR article !
OMG. “Common point of reporting for IT and OT is the CEO” (!!!!)
PS: this is the opposite of a coherent system — the two parts of a system that needs to work together are completely partitioned!!!! WOW!
Actually quite often those could co-exist isolated and also deliver 🙂 (within their area).
I have exactly that experience "How hard can it be?", it's only some MQTT data isn't it?
Fantastic diagrams — the zoom-in of complete non-overlaps of IT and OT are… you know… really funny and tragic. Kudos on making visible this problem to leaders who matter!
We are really happy to see that management teams all over the world are starting to use these diagrams. It's very rewarding and triggers us to keep on working on them!
Brilliant — the liaison becomes the first step to bridging the two silos, establishing an interface, the first step towards coherence.
Consistent theme: working together in multidisciplinary teams. Interesting to hear about manufacturing moving away from role silos, which is the context where role silos evolved. Competitive advantage is in the social domain (even more so with AI)
Absolutely ! Its extremely hard for an organization build on the foundation of Industry 2.0 to adapt to the current world
In the cross-functional teams, do you think they should have the people all the way from business to development? This is a tough topic in the manufacturing companies.
And while it's difficult to be able to involve the decision makers in different business lines, how would you organize the steering?
I think the crossfunctional teams can have business people, but honestly, I haven't seen it that often. Usually they prefer to work via steering commitees etc. Bringing them along will takes years of buiding trust but we've seen it happen.
Yeah, the issue I keep seeing is that the communication between the business end, POs, development teams etc. is lacking. And every layer in the middle is struggling to to create visibility and prioritazion and as a result, there is tons of fat in the e2e process. I have customer that just delivered AI data feature that was initially planned in 2021. And this is not an isolated case.
Reverse Ways of Working Maneuver. From Digital back into Manufacturing 🙂
:scales: And now the team from Legal & General, one of the UK’s largest financial services and asset management companies. @bharghava.bhogireddy2, Senior Operations Architect; @jennifer.pickard, Head of Engineering and @tariq.surty, Director of IT, are here to describe their initiative to overhaul the company's mainframe engineering and delivery systems.
Thanks @annp & @genek for the opportunity to present today. Please find attached a .pdf of the slides presented today. Many thanks
Oh, my. Wait until you find out why they called their initiative is called Project ImPala!!!! It is 😆😆
2 years as Britain’s most admired companies. Over 1T GBP in assets under management
that was a fun one, @willem.vanlammeren and @david856. Really reflects my own trajectory
I do a lot of work with embedded systems, also safety-critical. This means there's always this divide between doing things in "prototype mode" (not safety compliant, not mass-producable etc) versus "mass production mode" (do things properly, care for manufacturability, unit cost etc)
Am I managing? Well, it's not binary pass/fail, so there's always aspects I think could be better. But I like to think I'm making a difference.
Yes, same for me. The transitions are not working. But the Webcam stream is fine.
If you want to check out the cooperation models we found : https://itotinsider.substack.com/p/itot-cooperation-models. If you know of new ones let us know!
Hi Willem & David, Excellent talk - I am working in this space at the moment and we are in the middle of an innovation project to converge IT/OT at our company and would great to link up and collaborate on this.
Absolutely! Feel free to reach out via LinkedIn or email : https://itotinsider.substack.com/about
this what a live conference feels like sometimes, folks!
this is a wireless headset I think. Might be weird interference or codec issues. Darn software!
on a related note, everyone should read “wiring the winning organization” and what it’s like to live in the “danger zone”
THANK YOU ALL ! What a great group to be in 🙂 Make sure to subscribe at https://itotinsider.substack.com/ (PS: we started doing podcasts now as well to bring stories to the world, if you have any to share - we'd love to talk )
Slide transitions still not working properly. They get stuck "between slides". Anything I can do?
(Incidentally, I heard stories of incorrect JVM memory setting taking down an entire Z mainframe and all the workloads on it.)
(mainframe running in Azure? Oh, maybe the dev and test environments. Still amazing!)
“A Day in the Life of Leslie (the modern mainframe developer) Re-Imagined”.
Love the idea of Empathy driven. Curious as to how they did that. Empathy Interviews?
Some random thoughts thinking about this: 1. Start with understanding their context 2. People don’t care what you have to say until they know you care 3. Build relationships
Design Thinking Approach (Empathy based) - We have invited teams(devs, Test, Architects, Operations, Business Analysts,...) and started to ask questions and had two boards (what are your pain points, What good looks like/what do we want in new world).
2We had around 1200 -1500 post its. Which we still look at to make sure we have addressed all the pain points.
1I think the momentum grew as people began to see not just the range of issues but also similar issues other people were talking about. To take people away from their desks into a safe space and take a step back to look at frustrations & opportunities, We really tried to gauge it as this was their opportunity to have a voice and we wanted to keep that through out the delivery. And the promise of Pizza (and has continued) to help.
1Maybe I missed it, but is there any valuable reading material on Empathy related to this use-case?
✨ Please welcome @gayathrirgk, Agile Consultant (Unit Agile Leader), TCS and @mark.anning — Openreach Lead Agile Coach here to present, Charting the Course to Requirements Excellence in DevOps
🙂 Thank you @annp & @genek - for your convenience, our material also provided here.
4Thank you @christianhans.knuth. The slides will be available afterwards I believe if people missed anything. Thanks to the teams for posting the slides as we went through.
Can we blame it on hardware and then send it to the ops team? :rolling_on_the_floor_laughing:
Better to go full screen on the slides. A lot of screen real estate is wasted in a normal reading mode.
Problems escaped each phase of SDLC, lots of rework, long lead times (from concept to cash)
Feelings: frustration, finger pointing, across 500-700 people, puzzle doesn’t fit, no idea what full puzzle looks like
Goodhart's Law: When a measure becomes a target, it ceases to be a good measure
To really have fake change, also tie bonuses, promotion and pay to the targets!
the other angle might be premature convergence around targets leading to over constraining what is possible.. applying a deterministic mindset to an emergent space / domain... (complicated vs complex etc..) targets don't work well in the complex domain
someone shared this insightful post with me about goodhart’s law. i found some of the framing helpful. https://commoncog.com/goodharts-law-not-useful/
If (measure facilitates understanding/learning) 😄 If measure used to compare people/teams “performance” 👿
@jeff.gallimore this looks genuinely interesting - can't wait to read (hope it has the nuance around context and domain of a problem.. think Cynefin etc.)
there's an AI thought experiment might be called the paperclip maximiser that is quite interesting. https://en.wikipedia.org/wiki/Instrumental_convergence
Example: Wells Fargo retail customer account opening
Good talk! To me it's very important to think Specifications (BDD) and Requirements are different things.
Good talk! To me it's very important to think Specifications (BDD) and Requirements are different things.
Good catch! you can think of it in such a way that Specification means behaviour of the system and requirments tell the context or technical limitations
The best requirement is no requirement, but specification gives you the behaviour of the system witch is the key
Worst requirement that I have seen: "Autonomous Vehicle shall operate safely"
We have heard about SMART goals, what about SMART Specifications? Not giving the product team a clear goal to reach can end up in not delivering the value expected, losing time and resources.
GIVEN, WHEN, THEN. Is a solution for: • Capturing the dialog at the beginning • Preemptively find possible functional problems • Giving automatic acceptance criteria (even to the possible contracts) • Being the actual automated test cases at the same time • blocking possible scope creeps • Basically giving focus for the whole service and product development
This slide is fantastic — on what goes wrong during the co-creation process:
I’ve too often seen OKRs that are based on Say/Do outputs and other anti-patterns. I’m starting to just work on helping people use the https://www.soonersaferhappier.com/post/outcome-hypotheses-a-primer (w/o OKRs) as a first step
Also in the Miroverse: https://miro.com/miroverse/okr-canvas-sooner-safer-happier/
Loved this .Correctly depicted we need to both leading (present)and lagging indicators(post) .One gives feedback on the go more agile ,other after.
if you want to get into Gather, the conference’s virtual space, you can get there at https://itrevolution.com/gather
it’d be interesting @mark.anning to understand more about the love for co-pilot. Do you have data on this regard?
We started with a general awareness call for our Tribes, where the two us demonstrated Copilot's response to our prompts 'live' during our session. Gayathri followed this up with our Product Owner community, to help them craft their stories, and reduce time taken to do so. We're also developing a growing list of suggested prompts. Re data, we currently have much qualitative verbatim feedback from our POs which we'll then be using in conjunction with quantitative utilization data.
thanks @mark.anning, it’s really an interesting aspect to help product management to be more functional co-working with AI. Great job1 😻 I’m looking forward to see those prompts 🙏
I am really looking forward to Stephen Fishman and Matthew McLarty’s talk on optionality at the end of the day.
I am not familiar with their work so I don’t know whether by optionality they mean a custom pick & mix candy bag of business features on call through API’s, or whether their connotation of optionality is about postponing the moment your options become decisions (to extend the candy metaphor if I may: to decide what type of candy bar you want the moment you feel hungry instead of many days before).
I absolutely loved the graphic business novel about applying the concept of ‘Real Options’ in project risk management. The title of the novel is ‘Commitment’ (2013) and it is written by Olav Maassen and Chris Matts. See: https://commitment-thebook.com/. That book was more a traditional project management setting, I guess, and examples of applying optionality in a scaled agile setting are ‘Set-Based Design’ and ‘Thinking in Bets’ (Annie Duke).
Hi @j.hartman - we’re actually talking about both aspects (prep early to decide late AND enabling variety & combinations on demand)…. And…. Pirates!
Reminder: The talks are starting again in 5 minutes. Start navigating your way back to your browser. https://devopsenterprise.slack.com/files/UATE4LJ94/F06V49GSELX/timer.png
:flag-ch:And now, please welcome, from KPMG Switzerland, @vkonofaos — Head of Delivery Service Delivery Platform and @nandkishorgaikwad — Chief Architect - Technology. They will present an experience report on their strategic pursuit of Sustainable Business Agility, and the "SwissAlps" cloud platform that resulted from it, built upon Microsoft Azure.
One of the things I love about ETLS/DOES presentations are the breadth of industries covered, and how you learn about how these different types of organizations work — in this case, professional services firms, comprised of geographical entities, existing in a family of other partnerships.
agreed! Every industry is a tech industry
It was never explicitly mentioned, but decades ago, professional service firms were often notorious for not stellar backend systems — there were several examples in the 2000s of one of the Big Four consulting practices having material weaknesses, I think in their timekeeping systems, which were definitely in scope for SOX (SOX-404, as we called it back then.) Anyone remember that? So great to see better conditions to build apps to support the frontline consultants!!!
As someone who had to work in the handling of US clients in a European bank, I really would have preferred that you didn't remind me of that 😂
a departure from the genai-created images, i’m sure 😉 (also, a really good foundation for telling stories with flexibility to hit what’s most important)
I was today years old when I learned that Velux is a Danish company. Looking forward to the talk.
@nandkishorgaikwad what were the metrics and indicators you used to track progress toward the data center exit?
Hi @jeff.gallimore We measures the Actual servers number remaining, Overall RAM, Overall CPU and Overall Disk consumed of the Virtualized platform. The above metrics are monitored on monthly basis. At the end the actual number of Apps migrated from our CMDB.
1We don’t always need “the latest technologies” Boring Technology works well for most problems. Stop chasing shiny objects join The https://boringtechnology.club/
The following are some of my trigger phrases: • latest technology • Future proof (we will survive the implosion of the sun)
🪟 Please say hello to @henrikrhoegh, Senior Manager, Platform Engineering and Josefin Salomonsson, Director Digital Governance from VELUX. They will share how they embarked a new journey to better scale and optimise for flow.
I love this presentation, because they are so deliberate about thinking about organizational wiring — who talks to who, and why, about what. @henrikrhoegh Josefin Salomonsson
https://devopsenterprise.slack.com/archives/C06UL60MN1Y/p1714043304103429
“i made this diagram. it’s not scientific or anything.” :rolling_on_the_floor_laughing: 👏👏👏
Getting the abstractions right is the challenge. Brian Harvey on https://inst.eecs.berkeley.edu/~cs10/sp11/lec/01/2010-08-30-CS10-L01-BH-Abstraction.txt Brett Victor on https://worrydream.com/LadderOfAbstraction/
“YAML is the answer” hahaha. So good! (BTW, I recently heard podcast on inventor of YAML, and it totally changed my mind from ridiculing it to appreciating it. He just released YAMLscript, which I have complicated feelings about… …but it’s actually implemented in Clojure under the hood, which I do love. But configuration that is actually code can be scary)
https://www.therepl.net/episodes/52/ Super interesting, and not for the reasons I thought it would be!!!
It makes my job and the team's job easier (or not)! That's how I measure it!
What a brilliant framing of the source of cognitive load, and where to focus. NICE!
Platform engineering is not about engineering platforms. This is a conflation of product-centric or product-oriented delivery and platform engineering.
Here’s your kubernetes, helm, cluster, etc… “but it’s only YAML”. 😆
Anti Corruption Layer! Are you sure you really want to open that big yellow box? 😂
i love the idea of separating the developer domain and the platform domain. i see how that could really lower cognitive load. 💡
as long as the platform isn't overly constraining, Gregor Hohpe talks about this in his platform strategy book. Abstractions for platforms is really hard.. sometimes porous boundaries
That was the initial idea of platform "products" and platform teams. It's easy to lose your way though and end up just copying the abstrations of the lower layers "up", instead of adapting them for ease of use by the proiduct teams
@chriscombe this book? https://www.amazon.com/Software-Architect-Elevator-Redefining-Architects/dp/1492077542
@jeff.gallimore that is a classic read but this is the one I am referring to https://leanpub.com/platformstrategy
Platform engineering is the discipline of building and operating self-service internal developer platforms. Each layered platform is managed as a compelling product to optimize developer experience and productivity.
How do you plan 14 days if they have no estimates on how long work takes?
They don't plan them. They decide what is more important for the following 14 days.
No planning needed then. Switch to Kanban, stack rank/prioritize incoming work and take the top item each time. This will improve flow.
also what happens to stuff that would have gone into the backlog, do you ignore it and risk losing ideas or store it somewhere else?
I loved your Scrum like approach without a backlog, but a planning session to focus the team on what's important....very inspiring
@henrikrhoegh amazing IDP! I love your point on cognitive load ❤️ and it should be an interesting aspect that with the actual burnout data should be a paramount aspect of every CXO strategy. In the old days we were developing on a single language; end of discussion. Now, a dev needs to develop as a polyglot master, support the PO with BDD, the platform teams with yml/cdk/helms/etc. I guess your chart overlaps very well with mine I did with the Cynefin framework to refine the curriculum of the SW craftsmanship Dojo to create a sustainable cognitive load upskilling path. I’d like to discuss further your data and share mines.
Great presentation! @henrikrhoegh I didn't catch if Bura was something in-house, or an external product?
currently in house. we are working maybe open sourcing it. But it's super simple to make. Bura is inspired by Shuttle by Lunar. We just do more then what shuttle does. https://github.com/lunarway/shuttle
@henrikrhoegh great talk. We're going through this at Atom Bank currently so really resonated.
@henrikrhoegh Thank you for sharing! Great presentation, and really great achievements by the team. Approximately many developers are using the platform?
haha @sdc im told. Danish is easy to learn and impossible to pronounce. Only finish is harder they say 😛
Isn't Finnish the other way round – easy to pronounce and impossible to learn? But I am a Swedish speaker, I might not be neutral about Danish and Finnish 😃
😺 And now, let's welcome @cleng, SRE Engagements Product Area Lead at Google. He'll be presenting Overcoming Challenges for a Successful SRE Organizational Setup.
@henrikrhoegh you rock! (And we miss you at Eficode 😉 )
Our customer is all internal development teams that needs to run their application on our platform.
Many times customers don't love it when you can't give an estimate of when things will done. Especially if those customers have regulatory or contractual commitments.
“what do you do when things become scary?” Hahah. @cleng has talked about “haunted graveyards,” the scary services that everyone is scared by.
to the commentary on Product owning the what and why and developers owning the how.... i've be evolving my own thinking and am leaning towards product owning orchestrating the developers to the customers/users with the most impactful/biggest problems. slightly different twist that i've been exploring
Love this. One of the biggest things I learned in WWO book writing: “And the deeper the functional specialty, the more we need matrixed structures, where functional owners can define the WHO and HOW, and the product or value stream owner can own the WHAT and WHEN and WHY.”
this is a great frame, I recently drew a picture orthogonal to this but this gives me some new ideas for consideration ..
With healthy tension in balancing the WHAT and WHEN, so that not a feature factory.
“The role is a career dead end. Ambitious engineers will not bite.” (?!?)
There is no community, so they can’t grow — and thus, career dead-end, because there’s no path inside that product team! (Ah!!!)
isn’t it also the super-hero anti-pattern that will kill in laziness all the others?
Interesting — wired wrong, SREs lose sight of the business goals, become order takers, etc.
I think it can make sense as a temporary setup, SREs as enablers for the product team. But the product teams must get an ops mindset eventually, not always rely on embedded SREs
Agreed Stéphane, but as my grandmother likes to say, nothing is more permanent than a temporary solution 🙂
We actually do "embedded SRE" temporarily (for a few months/quarters) to have individual engineers deeply embedded in a critical project, working hand in hand with their Dev counterparts. But they still have their home team and will go back there fully when the project is done.
I fear the SRE’s become the doers of “grungy ops work” instead of individuals in multiple teams
“peanut-buttering SRE across to many engagement, when there’s insufficient coherence between them” — SRE cannot get deeper understanding to add value.
let the developers do their stuff!!!
“fungible” in this case is good? Meaning, the dev generalist can be deployed for more purposes to create business value?
this is why i am moving away from prescribing devs as just the 'how' and i'm pushing product thinking to be more about orchestrating developers to the users/customers with the biggest problems. I think the danger in saying product owns the why, what and when is in the 'what' statement. Product mgrs give devs stories and requirements and don't give them access to the actual problem first hand from the users
I agree. Developers need to know a bit about everything. But they need specialists like product, SRE, etc. as force multipliers. Developers should be a "fungible resource" in the sense that they can flexibly shift their focus.
(Regarding the cats: “does he remind you of a James Bond villian?” someone said, after @cleng said “you only understand a system once you see it on fire… in production…“. ). 😆😆😆
“Training Is Useful, But There Is No Substitute For Experience.”, Colonel Rosa Klebb, From Russia With Love
I was planning to stroke a cat while sharing my plans with you and the rest of the world from my secret base... 😁
Just going back to SRE struggling with a bigger picture, there’s an old saying “when you’re up to your backside in alligators it’s difficult to remember the original plan was to drain the swamp”. If your SRE’s are drowning in incidents then they’ll struggle to see wider business goals.
@cleng Wasn't there a mention earlier that SREs have a high work satisfaction – was that in @jpetoff's talk? It would be interesting to see where the difference of perception is. I tend to agree with Christof's view.
Antipatterns are things that CAN happen. They are not a reality everywhere. It would say that antipattern #4 is actually the easiest to avoid / fix. But I have seen it happening and it can be disastrous for the success of production improvements.
Oh, my. “When SREs makes production more complicated than it needs to me… not necessary with malicious intent.” I think that’s how it feels when someone says to dev, “here’s the kubernetes platform you need to run your container in. good luck!” (Similar to what @henrikrhoegh from VELUX described in the last talk)
My goal has always been to work myself out of a job. I mentioned that in an interview and saw looks of horror on people’s faces. Needless to say I didn’t take that job.
We actually have a saying that you should automate yourself out of your current job every 18 months... because if you don't, you're not really adding value and will eventually drown in ops work. However, that doesn't mean you will be out of "job", it means you're ready for the next "job". There's always the next, bigger problem to solve.
Would there be any value in rotating some people into and out of SRE roles so that they get a wider picture?
Totally! We have a program called "mission control" where software engineers do a 6-month rotation in SRE. They get all the training new SREs get, do the same project work, and go oncall for our production services. Afterwards, they can go back to their old role, or apply for a role in SRE.
"complex infrastructure that only can maintain": how to keep a lid an implicit incentive for people to make themselves indispensable (not only SRE)?
I think it’s not necessarily “people making themselves indispensable” it’s also people struggling with determining proper abstractions
people need a vision of the future; you need to paint an attractive picture for them about what would improve for them if that complex infrastructure only they can maintain disappeared
Doesn’t it all begins and ends with consensus (or lack of) on “you build it, you run it?”
Yes. That's a fundamental change in the dynamics. But most of the points in this talk are, I believe, orthogonal to whether there is someone outside of Dev who runs production (could be an ops team that's separate from SRE).
there is a wide spectrum between "you build it, ops run it" and "you build it, you run it" - this is the space to explore in your particular org
Well, I'm sure you get them less wrong than last year. Maybe you can continue being less wrong about them? 📈
Are these anti-patterns hurting you? Are the ideas from my talk helpful to mitigate them?
✅ Please take a moment to answer these two questions to help us improve the programming: https://forms.gle/frVdffNr7pYdNfmW6
@cleng amazing talk, and lovely anti-patterns list 😻. It’d be fantastic if you have some measurement to make them objective to facilitate the discussion with leadership flipping difficult conversation from perspective to data. If not; it would be amazing to create some
Same issue with DevEx as well – how can we sell the value of reducing cognitive load for developers?
that one I guess is easier. Gallup report on engagement, plus infinite potential burnout report, combined with McKinsey report on upskilled dev efficiency VS disengaged unskilled ones. DevX has more data, at least that I’m aware of, but Chris topic is an unexplored world (for me at least)
Shameless plug, but at Gartner we have published extensively on the subject and we would be happy to talk to your leadership. Example publications: • How to Build Executive Support for Developer Experience Initiatives • The State of Developer Experience Initiatives: Benchmarking Data From the Field Map the developer journey, pilot, connect to org priorities, manage the messaging to get buy-in. We've got case studies and testimonials as well to support clients.
Yeah, I agree that data would be very helpful and that there isn't much in this space yet. My team is actually exploring some ideas in that space, but it's too early to talk about them yet. If anyone here has data (or ideas what to measure), I would be very interested to hear about it!
The problem I see from people setting salaries is often a feeling of "these devs get paid better than us, and they are complaining that their work is too difficult"
I’ll try to connect with you @cleng with the data I’ve and the ideas I made with them. I’m really curious to understand if we find some interesting behavioral engineering patterns
If you pay your software engineers in SRE worse than your software engineers in Dev, you have a big problem.
I agree – when SREs are seen as "just ops", it often becomes the case...
Great talk @cleng ! I do see parallels to your antipatterns in my org - I see our Rely/SRE team is behaving like an oncall/ops team.
Reminder: The talks are starting again in 5 minutes. Start navigating your way back to your browser. https://devopsenterprise.slack.com/files/UATE4LJ94/F06V49GSELX/timer.png
⚡ Time to have some fun with Lightning Talks! Please welcome: • @michele.brissoni856, Independent SW Craftsmanship & DevOps Coach • @nathen.harvey, DORA Lead and Developer Advocate, Google Cloud • @sascha.schaerich, Lead DevOps Evangelist, Deutsche Telekom IT • @marceloancelmo, Head of Solution Architecture, KPMG Swizterland • Neru Obhrai, Neru Obhrai, Principal Consultant, Radtac
1Welcome to my virtual office, and where we run all our dojos. We can’t be more happy about Gather Town and the gamification effect it produces working with devs
yeah at that time before I refined the idea of the dojo it was indeed a scary monster. Happy that the presentation transfers that feeling 🙂
@michele.brissoni856 out of curiosity which text to image tool did you use?
Ciao @jonathansmart1, all the details are in my https://www.linkedin.com/pulse/behind-scenes-revolution-tech-leadership-summit-michele-brissoni-pkdtf/?trackingId=xQcYNnNFRLywXzqc7F7ssQ%3D%3D in the news letter the forge of unicorns. short answer: I experimented with DaVinci, ChatGPT, MS Designer
Wow! "Once the script was locked in, it was time to take it to the next level. I decided to explore http://play.ht, an AI text-to-speech service, to create a voice clone of myself, ensuring it could flawlessly deliver the talk within the allotted time. Cloning my voice? Now, that's some next-level tech wizardry! 🧙 Especially considering my dyslexia! 🤐 And hey, if you're wondering, yes, I even applied a nerdy test-first approach to preparing my conference!"
@jonathansmart1 the perversion to be a TDD coach is addictive 😛 so I had to test my talk before doing it 😄
1OKR at board level: North Stars. Neuroscience & Behavioural Psychology. Nice.
OKR combined with behavior engineering can be really powerful. I guess it aligns pretty well with the BVSSH framework of OKR as hypotheses on outcomes.
The image of the dojos on the staircase was awesome — didn’t even think things like that could be done with text-to-image!
AI is changing the game… if you know what you wanna achieve now the possibility co-working with AI are simply astonishing.
PSA: it took me an embarrassingly long time to figure out how to access Gemini 1.5 — here’s the link. IIUC, API access is free until may 5 or so. https://gemini.google.com/app
Encourage your team to reflect on how they’re doing today. This year’s DORA survey is open now and includes expanded exploration of AI, platform engineering, and developer experiece. Take the survey and discuss the insights you’ve gained in your next team retro - https://dora.dev/survey
Thanks @michele.brissoni856 and @nathen.harvey! And now 5.5 lessons from @sascha.schaerich!!!! 🎉
1“it’s more about culture and mindset than technology” also yyyyesssss…
In the end of course, but at least in our company we started out focusing a lot on the Technology-Stuff, maybe because that is easier and quicker to understand?
Totally understand this. I notice that here too. Technology is always there and the first look is on the closest. With the right questions, we can broaden our perspective, which leads to better conversations and better outcomes.
I'm working with a company that tried changing their ways of working several ways / times and every time they ended up being up each others throats (2000 employees), now we have worked with them on a transformation in terms of processes, mindset and tooling. 3 years later we are seeing finally the culture change. Rewarding..... but ohhhh so slow 😬
We need to give people the opportunity to unlearn what they have learned in the past - everyone has a backpack - and then learn new things. 😉
“especially for German companies” “especially for companies like us”. 😆
What's the difference between English punctuality and German punctuality? (I will permit myself that one 😂)
So true! We recently did a MagentaExchange with 14 people from Germany and Europe, and 14 people from T-Mobile US, and the difference was noticeable, e.g. when getting everyone on a bus!
Not a joke, the color draining is real… ;-)
@sascha.schaerich, we need more bald guys with long beard! 😄 👏
I stopped delivering DevOps trainings at Deutsche Telekom in 2019 or so, so I was able to rescue some colour in my beard. It's sort of salt-and-pepper now 😄
So you got out just in time apparently… ;-)
What fantastic tips on mentors and mentees, codified at KPMG Switzerland! @marceloancelmo
Glad you liked it @genek. We got a lot of positive feedback after we put the mentees on the driving seat on our mentoring program
you can declare what you want to learn, or what you want to teach — nice!
I’ll be honest, I’ve heard about neurodiversity, and mostly thought of it in context of children and school — its implications for the workspace was startling and sobering. Which is what made Neru’s talk so memorable!
You know, I have a neurodivergent kid, and I suspect I'm neurodivergent myself -- but I've never actively considered how this topic applied to my coworkers -- or indeed to myself. I'm kinda baffled at myself right now.
Great topic. Too many cursors on a Miro or Mural board is something I've had feedback on as not being good for neurodivergence.
Thank you @neru for sharing and Gene for including this talk!
I'm so impressed with the breadth and variety of topics covered by this conference
really touching talk, as a neurodiverse guy (ahhh BTW Mike in the presentation, it wasn’t me 😄)
⛵ Up next is @lalle, founder of a company called Scling that does data engineering. He's here to teach us something that is relevant to every organization in his session, Industrialised Data - The Key to AI Success.
This picture has a nice visual on neurodivergence in the workplace. It is in Dutch however, but it can easily be translated with Google translate or Deepl.
It contains 4 types of neurodivergence with so-called ‘spiky profiles’: peak strengths and valley weaknesses.
Very interesting, thank you. And trying to understand the Dutch makes it even more fun 🙂
It should not be that hard. One of the four profiles is exactly the same abbreviation in English and two of the other three profiles only differ one letter. 😃
In education, it's interesting to explore how AI can help teachers scale their lessons for neuro divergent learners by adapting lesson plans into multiple delivery formats almost on the fly. Like giving every learner their own personal teaching assistant that can adapt to their specific learning needs. No reason for that to be contained to just children's learning environments
“Leader vs. Rear” <==> “First vs. Worst” <==> “high vs. low performer”
It was interesting working with Dr. Steve Spear — in manufacturing, it’d be amazing to see an 8x difference between high vs. low. In tech, 3 orders of magnitude difference!
billions of datasets per day: Google 2016. It is joked that all Google is are a gazillion of microservices with protobufs, deserializing, transforming, serializing, and then sending to another microservice.
Spotify: Discover Weekly: just 3 engineers, 3 weeks; one of the most important services in Spotify history
Imagine a world where that app team could just connect to the right relevant data set, connect it to some business logic, a button. With no approvals, coordination cost, etc…
“The car is full of disappointments” :melting_face:
TBH it is a magnificent car in most aspects, but the digital experience is missing so many opportunities. Which can be said for most cars. If I had known what I know today when I bought it, I would still have bought the same car. Even if I had infinite budget. So they do some things right. 🙂
That's exactly it, make the telemetry work for the customer and not against them!
I don't think it's because the devs don't have access to the data. These are product management decisions.
it is always easy to blame product, engineers are smart they can influence outcomes too
I was half listening to the lightning talks because I was trying and fighting to get my engineers access to basic datasets!
I have spoken to devs and managers at Volvo, and they confirm that they don't have access to data. In order to do something with data, they have to fill in an application and get it approved by HQ. I once visited a bank that had not one, but two committee decisions before you could start writing code for a data pipeline.
Vehicles are not seen and made as a complete product. They are many parts bolted on. Except Tesla perhaps, where they started with the software and built a car around it.
Indeed. The car manufacturers essentially ship their org chart as a product experience. This clip with Ford's CEO is very insightful: https://youtu.be/HrNN6goQe50?si=dwUVUa1wl2GEu-GG
Except for firmware drivers crashing, and randomly rebooting the displays and safety critical systems while driving. Even Tesla!
This is so critical to digital twin and value stream management efforts
“what’s missing is all the data plumbing, which enables people working with the data to do their work easily and well, and enables independence of action”
“what’s missing is all the data plumbing, which enables people working with the data to do their work easily and well, and enables independence of action”
It's the same as in technical products in general, or indeed as in organisations. The interfaces are what's hard.
the same with people - it is interactions that is hard (often between teams) but also within
... and they are easy to overlook. People see the objects much better than their relations
True indeed. The dynamics with data pipelines are somewhat different than for e.g. integration with microservices, since there is often wide coupling in the business logic. For microservices, each service typically has a few neighbours that it interacts with, and the coupling is small. In contrast, data pipelines that e.g. compute financial risk or detect fraudulent behaviour can consume data from across the company.
The data mesh is an organisational pattern for scalability. It can be applied in both traditional data warehousing environment as well as in "industrial" data lake environments. A data mesh can work well to address scaling limitations, if a foundation of data democratisation is already in place. Most data mesh implementations that I see are in environments that are far from hitting scaling problems, where data democratisation has not been achieved, and they in practice end up enforcing existing data silos.
1My father never understood what I did and used to say “Chris pushes buttons for a living”. I mean technically he wasn’t wrong but… 😄
🍔 And now, let's welcome @mik, CTO of Planview and author of Project to Product. He's here to speak with Gene about Decoupling from Foundational LLMs.
Thanks everyone for listening! I'll put up the slides online and post a link here. Please reach out to me at <mailto:lalle@scling.com|lalle@scling.com> or on LinkedIn if you have questions. Also, if you want more information on the "industrial" style of data engineering that I preach, have a look at https://www.scling.com/reading-list/ and https://www.scling.com/presentations/.
My slides are now online at https://www.slideshare.net/slideshow/industrialised-data-the-key-to-ai-successpdf/267534646
50 teams learning LLM intricacies and implementation means lots of separate parallel undifferentiated efforts
…unless there is a shared consciousness, coordination, and intentionality on the decentralized experimentation. then it can be a major accelerator. but that’s really hard to do…
Wow. I hadn’t thought about the side effects on coupling of helping others learn prompt engineering.
Coupling the wrong way. Not modularising. 24 people on centralised core re: LLMs (EAs?)
meaning the data scientists will use AI to write the AI tests for the AI models? 😉 what could go wrong
Isn't this the same as developer writing unit tests that somehow always pass whether they should or not 😄
yup, eh ability to write tests, with the raise of AI starts IMHO to become a must to have skill not only as a developer. if we trust AI without having proofs it does what was the behaviour we requested to get help with, we can get out of control of a dangerous dog
Emergent. Have a set of Principles (Principle led, apply across contexts)
Biggest lesson: have conversations with architects and product leaders on product teams and AI teams (i.e. collaboration across disciplines)
Low switching cost for LLMs. Don't know how they will evolve.
abstraction is quite hard in that case... knowing where the "puck will go" or your APIs keep breaking.. let alone your teams and interactions
Optionality. Centralised to lower switching costs in the future.
“Don’t know how they will evolve.” but we do know how fast they are evolving… which is :lightning: fast. makes this approach even more important.
I really like the focus on continuing to try things and learn as far as how folks work together and learn together and shifting when learning and data proves the need to. And not using new tech everywhere simply because it's new and cool but with so much strategic intent and connection to business needs and paired with a focus on decentralizing. Nailing it.
Woot! Woot! So much goodness being shared on optionality! Can't wait for my closing talk with @mattmclartybc on creating scaled optionality
The talk or the book? We did have good news last week that our launch date moved up to September 🎉🕺
If decentralised, switching cost would be high [trading short term, medium term and long term]
So is the core AI team at Planview a Platform team or an Enabling team (as per team topologies)?
I am guessing there is more than one team and platform teams are actually mean to be teams of teams according to Matthew and Manuel.. so it shoudl be a team of teams offering a platform that includes enabling, complicated subsystem and platforming etc..
then the stream aligned teams would be consuming the platforms through x-as-a-s (hopefully) which having an enabling team facilitate the onboarding and collaboration for future building / growth ..
if that model doesn't work the other teams will build their own to move more quickly so collab is critical
Basically unifying policies, but decentralizing implementation to suit team preferences, or technology innovations. That's how you can avoid getting locked into old technology, or bleeding at the edge of it.
This is the same challenge as early DevOps: DevOps engineer on every product team motivated the rise of platform team investments
especially if you dont have a way to visualise your cross functional teams and their interactions through out an org
@stevesargon, yes org debt > tech debt, but that addressing it might be considered "boiling the ocean." Plus, politically it's not as likely to be rewarded in the short or medium term.
Also good point about there being no short term rewards. “It’s hard to get people to understand something whose job depends on not understanding it” - forget where I got this quote
@stevesargon, as long as someone has his or her eyes wide open that there's risk and is comfortable embracing it, I say go for it. I'm something of a contrarian myself, always optimistic that there's an opportunity, but it's not for the weak.
https://leaddev.com/process/ai-governance-policy-engineering-managers-needed-yesterday
🏦 Up next is @sdc, Platform Engineer at DKB, the second-largest direct bank in Germany, here to present Introducing a Product-based Approach in an Engineering-focused Container Platform Team.
Starbucks Nitro Cold Brew. Though it is not for the faint of heart. I tried my first one, had about 1/4 and felt like I was either going to have a heart attack or sprint somewhere. ⚡
staging environment? with a change board for talkers to move to live? (kidding )
The thing I like about the events with lots of engineers is that when there's a problem, there's about 20 different ways presented on how to resolve them 😄
2Drinking more than one of these has some interesting side effects. At minimum, it does make you feel alive! 😂 😂
A live event is not enough?! :rolling_on_the_floor_laughing:
See my note in another thread. I felt like I was going to have a cardiac event or sprint somewhere.
Too industrial 🙂 I am powered by: https://www.bobolinkcoffee.com/Coffee.asp?op=TheBobolinkCoffee today/
BTW, speaking of tech: I observeds several times today that my video stream was such low quality that I couldn't read the slides. No idea what to do about taht other than to change to a different platform
Well, that's the confusing thing. I'm in Munich too, begind a GBit connection, so there's no reason other than the platform being somehow funny 🤷
“communication often was focused on [arcane apsects of infrastructure] engineering, which was not helpful to people without infrastructure background”
“which VPC, which EC2 instance.” (and other stuff devs don’t care about)
“logging” (and other stuff devs don’t care about) 😆 My week learning about Java logging was on the one hand super useful, because I finally understood what goes where, but OTOH, I got utterly no satisfaction or joy out of it, because I didn’t actually care — I just wanted to see my logging statements!
Devs should care about what to log and how to log, but not the intricate infrastructure details
i.e., I really didn’t care. I just want to see my #$@! logs.
It’s another kind of decoupling, you want to get your logs without worrying about which mechanism is providing them
It seems a lot of companies are building DevX and platforms and it seems there are many commonalities Just like k8s arose from a common need do you see a common platform)DevX open source offering arise?
that is why things like backstage exist to not reinvent the wheel especially where it is non-differntiating for a company
but form a wardley map perspective many of these topics are still quite early so over time they will converge etc.. we see DX already happening there e.g. DX, LinearB, Jellyfish and many more..
Precisely there must be a higher level of abstraction that we can all benefit from
that is precisely what backstage tries to do with its plugin architecture
To me, Jellyfish is more what we call "Engineering Ops". It serves management more than devs. Developers really don't want to be measured on their productivity.
Love it. To tie this to @mik talk: platform is evolving — I want to be completely indifferent to the logging framework. Tell me what to do so I can see my logs. Change it as much as you want!
@sdc How you were able to understand what is / is not clear for the platform? surveys?
Is your question related to the heat map? In that case, I meant how well-defined the capabilities are for the platform team itself.
We are not using surveys, we rely on direct contact with product teams for the moment.
“people not used to thinking so abstractly about the service” — often cared too much about the implementation details.
It’s not just understanding/communicating value, it’s being able to answer “How will we know if we are adding value?”
@mik, as far back as 2017 I came across TaskTop and shortlisted it for a client project! It's great to see how good products continue to evolve.
interesting to see risk as a lever for value. - that is a nice frame
It’s too easy to get sucked into and enamored with the technology you are using/building and lose site of the problems we need to solve. Problems not just for our customers, but our customers customers.
getting people to do anything is challenging.. invite over inflict !
@stevesargon have you looked at John Seddon's 'Check' model for this? Great way to measure and explore a system, expose the conditions, assumptions and thinking that is driving current problems and then gain buy in to solve them end to end
It was designed for use in customer service operations but I feel translates well into software delivery with a few tweaks, happy to discuss
@chris.atkinson funny you should mention Seddon, my book club recently finished https://beyondcommandandcontrol.com/beyond-command-and-control-book/
key learning for me in that was failure demand vs value demand.. your contact centres can play a huge role to learning from customers
yes – this is actually one thing I had in mind when saying that everybody should have a product mindset. This kind of feedback is difficult to see "from the inside" otherwise.
@chriscombe they really are your eyes and ears when it comes to gaining real customer insight
@sdc I have witnessed the complete relaunch of DKB user interface (Browser and app) as a customer and am really impressed. The bank switched from a very old-school interface to a modern and easy to use interface within few releases and it still gets better all the time. Probably the platform improvements made all of this possible, so thank you for that and for the interesting talk. 🙂
Thanks for your feedback! The particular platform I was talking about wasn't involved though, so we unfortunately cannot take credit for that 🙂
https://a.co/d/2a8qdYq The Async-First Playbook: Remote Collaboration Techniques for Agile Software Teams by [Sumeet Gayathri Moghe]
Just have to love ETLS… reading list has grown exponentially last 48 hours 😁
My working list in no particular order 1. Co-intelligence: living and working with AI 2. The async-first playbook 3. Platform Engineering Maturity model - https://tag-app-delivery.cncf.io/whitepapers/platform-eng-maturity-model/ 4. Team topologies 5. Transformed by Marty Cagan
Does @genek & co provide a reading list after the event? That would be awesome!
Here’s a (not super) photo of a sticker we had a our Las Vegas Summit last year. It is our IT Revolution Starter Pack of books, Keep in mind it doesn’t include new titles just released in 2024!
*This is also not a full listing of the IT Revolution library. 🙂
https://devopsenterprise.slack.com/archives/C06UL60MN1Y/p1714043304103429
1Reminder: The final talks of the Summit are starting in 5 minutes. Start navigating your way back to your browser. https://devopsenterprise.slack.com/files/UATE4LJ94/F06V49GSELX/timer.png
🎬 And now, please welcome @bcipot Senior Security Engineer at Synopsys. He's here to present How Cyber Threats Can Turn into an Unwanted Real-life Hollywood Blockbuster, teaching us about the epic level of cybersecurity issues that every modern technology organization faces.
https://danielmiessler.com/ Cannot recommend Daniel's work enough! His Unsupervised Learning Newsletter/podcast is really concise and very informative
I really love the talks we have had in Gather today. So many interesting, big and relevant topics! 💛 🙏
One of the most remarkable parts of the story: the Debian build servers are taken down so they can be rebuilt from scratch, presumably because they were assumed to have been potentially compromised. (!!!!)
Fantastic detailed writeup of the XZ timeline: https://research.swtch.com/xz-timeline
Interestingly, my unrewarding week learning about Java logging was 30 days before the Log4J event. 😆
…and what if OpenSSH wasn’t actually the real target? https://twitter.com/RealGeneKim/status/1782460738570944709
BTW, while writing the WWO book, I became addicted to the “Darknet Diaries” podcast — the episode on markets for 0-days, some of the ransomware industrial complex, was so good. https://darknetdiaries.com/
Some faves: • Ep 126: REvil (ransomware complex) • Ep 123: Newswire hacks (to get early access to earnings reports) • Ep _: Click Here “Lapsus$” • Ep 100: NSO: spyware • Ep 98: Zero Day Brokers (!!!) • Ep 73: WannaCry • Ep 54: NotPetya (mentioned by @jonathansmart1 • Ep 53: Shadow Brokers (about the NSA toolchain)
Just to name a few. 😆 I gained so much appreciation about the vast economies that power the zero-day marketplace, ransomware, and what nation-states do.
How many transitive dependencies to integrate instagram and slack —
With everything that needs to be integrated, automation is a never-ending journey.
I was reading lately that EU commission decided to force the SBOM for security issue indeed to evaluate dependencies injection
📦 Up next is @jacob.brank, Lead Product Manager at Forto and @andy.duncan, Head of Engineering at Equal Experts. They're here to present: How Forto Solved Their Global Freight Trading Strategy Problem
Interesting trivia fact: I met @andy.duncan in 2010 at Mountain View DevOpsDays run by @patrick.debois256 @jwillis @damon!
Was that the one in the really strange warehouse/garage sort of location? If so, I was there and met @patrick.debois256 and @jwillis there for the first time as well. I was one of maybe 5-8 women. I think it was Patrick who found the women and got us all together for lunch at a picnic table, haha.
I think we were from IBM, Etsy, Netflix, Chef, and... somewhere else.
Patrick was talking about community building and John was talking about Deming. 😉
I believe it was https://legacy.devopsdays.org/events/2010-us. My strongest recollection was being blown away hearing @allspaw for the first time. And then a casual pizza with the devs from IMVU, talking about Continuous Deployment. 🙂
Argh, I'm sorry I couldn't make that talk - but I hope the canonical comic got a mention - I made a slight variation: https://infosec.exchange/@beny23/112234581994347708
Check out #C04ED43AQAC — all messages tagged with bookmark emoji gets copied there, which people have tagging! (@jeff.gallimore: we should include that in the “instruction manual for the conference”)
Freight forwarder own no ships — that’s the capital intensive business
I learned that most of the actually shippers are privately owned — a fascinating business!!!!
plus teletypes, faxes and paper originals sent by couriers worldwide... )
This is mind blowing to me in 2024. And it happens even in our own company... You know, the one known for automation and DevOps. I think it just shows how far from the ideal the world still is
imagine what might happen when freight was promised to get somewhere, but it didn’t make it on the freighter, and the firefighting that would happen to get it to where it needed to go, given very finite shipping capacity…
Shipping is a fascinating business with a lot of long term thinking. 25 years ago they used to close deals on Yahoo Messenger. Trust is a major thing. OTOH, they are slow to adopt new tech. They are quite practical. If it works, don't change it.
Matching supply and demand — in energy markets, matching supply and time
i really love it when constructs and concepts from completely different domains make it into technology.
Totally agree! With the companies I work with, I constantly see the same patterns repeating regardless of the industry. And often times those companies have convinced themselves that it's an issue specific to their company / field
I was working on multi-modal freight software 25 years ago and got made redundant! I wonder if they'll ask me back now 😂
Love the problem space => solution space framing. Did you have to validate assumptions early on?
in another version of the presentation, I talked about trying to break the model over and over again.
a bit like TDD, find a scenario that you don’t think will work, rather than looking for more scenarios to reinforce your current position.
so we tried to find all the crazy combinations of ports and cargo-types, and etc. that forced us to adapt the model.
Outcomes: increased gross profit! 30% increase in productivity of allocation -> booking
And pricing based on the combination of To/From/Client/Line/Cargo type etc. is just another yet super complicated issue 🙂
And here's another one multi-nodal mobility solutions! Also been working on that one!
Thanks for the opportunity to share our story. I hope it encourages some folks to pursue simple solutions to even complex-appearing problems.
📚 And now, bringing us to a close is @mattmclartybc and @internettitan, here to give us a preview of their upcoming book, Unbundling the Enterprise - APIs, Optionality, and the Science of Happy Accidents.
…and I met @internettitan in 2012, even before Phoenix Project came out, and SxSW!
i never thought about it before... but our meeting was also a happy accident... I chose "When IT Says No" as a filler session before the one I wanted to attend (that I no longer remember ... no recollection of the name or subject).. I'm not even sure I attended anything else that day. Gene's tale of the schism between IT and Business was so visceral and on-target, I was hooked. It was exactly what I experienced in my role leading dev at Autotrader.
Blackbeard would also find treasure on the high seas by plundering other ships, matey!
The Steve Yegge Rant!!! https://gist.github.com/chitchcock/1281611
The rant: https://courses.cs.washington.edu/courses/cse452/23wi/papers/yegge-platform-rant.html
Steve’s video explaining the whole thing: https://youtu.be/6GL7gykr1ZE?si=xkY-SLf-zIbJkQSy
The term I learned with Steve Spear: “because they regained independence of action!”
i.e., AWS became a module in an of it self, to enable teams to work independently, and then eventually sold externally, becoming the largest contributor to profit in the company.
APIs are just one way to reduce switching cost — changes can be made on just one side of the interface!
Independence of action <==> one side can change without coordinating with the other side.
a tale of 2 pirates... Greybeard (master of the old ways) and Captain Bob who lives in the land of 1,000 shovels
this one almost made it into the book.... we have a plan to surface it when we get closer to september launch
how can you have a book/talk about happy accidents without the GOAT?... All hail Bob!
“when looking at colleges or future mate, do you decide or commit early, or do you do this late? Late, because you have more information”
“Oops” reminds me of <Failure:%20Why%20Science%20Is%20so%20Successful%20Stuart%20Firestein|Failure: Why Science Is so Successful - Stuart Firestein> great book
Google Maps monetization case study was awesome — I had no idea!
I’ve used value exchange maps for 2 decades — until their book, I didn’t actually know where it came from!
@internettitan @mattmclartybc …and I can’t remember what body of knowledge it came from. HELP! 🙂
Comes in many forms, but I think Roel and Jaap deserve special credit: https://www.thevalueengineers.nl/
David Bland's "Start with Learn" resonates with me a lot lately: https://www.precoil.com/articles/start-with-learn
After this trailer, how can we be expected to wait until September for the book!?
This was a very enjoyable and high quality event with tons of information to process. Thanks you!!!
Thank you @genek @jeff.gallimore @annp @mvk842 @alexb!!!
1Hi Tempo, massive amount of inspiration and practice-based examples... it was really good
Those quick two questions to help us improve the programming! https://forms.gle/WxkDzzLdavoT1ZRF7
Yes congrats to the boss wife and everyone else on the planning team!
1Thank you! has been an amazing learning experience and I have many leads 🙂 I will spread the world
2After drinking from a fire hydrant, things feel so quiet when DOES // ETLS wraps up!!!
The tragedy of the summit is when it ends and the great lull begins… Still wondering how to successfully keep the community communication and energy going between events… Thoughts everyone?
Thank you all! and thank you for at least tolerating my "self-deprecating" sense of humor (as I call it)! 😂
Thanks again for organizing the watch party- made all the difference for me!
I'm mostly amused because sometimes I still need to resort to calling it "The conference that used to be called DevOps Enterporise Summit", and I'm glad I'm not alone in tripping over it. But at the end of two long days, you're forgiven for any and all verbal slip-ups. You led us through the conference masterfully.
At times I feel like my primary job is a video recommendation engine. 🙂 Not a joke!
You and the team curate and set the standards for what we get to experience…. 💕
More watch parties! I want one near me. Perhaps connect with DevOpsDays organizers to help coordinate/host (fyi I’m one for NYC)
@stevesargon, let's do a New Jersey watch party next time. There are plenty of us in New Jersey.
1If we have one of these in the U.S. time zones, I bet Red Hat would host a watch party in Raleigh.
1I noticed so many people in Munich. We must have a watch party next time
We must have a mini-DOES get-together, either at a "serious" location, or at least at a beer garden or something
I'm serious. Any Münchners (or people visiting) reading this, I'm happy to serve as point of contact to organise something. Find me on linkedin or write to me at <mailto:luca@ingianni.eu|luca@ingianni.eu> and express interest, and I'll try to figure something out!
next time I’m gonna be in Münich, I’ll ping you Luca 😉
Good conference. Thank you for the learning and the interactions. Until the next one! 🙂
3Thanks to the organising team, also to all of you in the chat who made the conference even more awesome!
What a great format of a conference! Thanks so much to all organising it and presenting their stories! Very inspiring!
1👍 This was my first summit, great experience and some truly valuable learnings which my colleagues and I can bring back into our organisation. Thank you to everyone who made it possible.
4Reminder: Have you been thinking of someone you know who needs to see talks like these? We have videos from all our past events in our video library so go ahead and share! The people you share the videos with just need to enter their email address to start a free trial. Individual (unlimited views for one person) and Organization Memberships (unlimited views for an entire organization) are available at https://itrevolution.com/upgrade. https://devopsenterprise.slack.com/files/UATE4LJ94/F06VC7P5Y5U/videolibrary.png
Thank you everyone - past two days have been awesome! Looking fw to next edition! 🙏 🙏
4@mik If you're still watching the chat... I'm really curious how you centralized prompt engineering into one repository. Can you explain in a little more detail how that's done? We'll probably want to repeat that pattern.
you could take inspiration of the langchain hub model - https://github.com/hwchase17/langchain-hub