Excuse the navel-gazing - “TikTok heavily favors videos edited inside their app. YouTube is mainly a thumbnails game. And who knows what Instagram wants."
The AI Trust Fall - Even though humans are more error prone than computers, we trust the work of humans more…perhaps because we understand their errors much better and how to debug them. You know, because we’re humans ourself. In contrast, with AI’s, we have no idea what’s going on nor how to fix errors. This feels like some mystic, Talebian wisdom like the benefits of getting Fat Tony to eat more humus and drink more wine for long life.
JPMorgan ramps up prompt engineering training, AI projects | CIO Dive www.ciodive.com/news/jpmo… AI to replace executive jobs: “Another AI tool allows employees to query a model meant to respond as Michael Cembalest, current chairman of market and investment strategy within the asset management division.” // With the insane comp. that executives get, maybe AI would be a lot cheaper!
Training is not the same as chatting: ChatGPT and other LLMs don’t remember everything you say - “Every time you start a new chat conversation, you clear the slate. Each conversation is an entirely new sequence, carried out entirely independently of previous conversations from both yourself and other users."
HUGE Google Search document leak reveals inner workings of ranking algorithm - Reverse engineering Google search ranking based on method calls from leaked code.
Interview - WSO2 CEO Sanjiva Weerawarana on being acquired by major investor - “In the last financial year [CY2023?], WSO2 achieved revenues of $100 million”
Survey reveals generative AI employee fear - Lots of hopes and dream, not enough projects and experiments: “IBM’s research found that less than half of organisations are focused on GenAI pilots – and another 24% are doing nothing at all. But almost half (49%) of the CEOs polled expect to use GenAI to drive growth by 2026. According to IBM, this is very ambitious, as only purposeful transformation will make it possible."
This week, at my work, we launched the Tanzu Platform. This is both actual, new things and strategic tuning of many of our existing products. In brief, we’re focusing a lot more on private cloud PaaS now, and private cloud data services. (Sure, you can do public cloud stuff too, and our customers tend to both, with a lot of private cloud.)
The Tanzu Platform needs some kind of IaaS “dial tone” (either VMware or Kubernetes) and then you can run a multi-tenant, global PaaS, get integrated DevOps/platform engineering tooling, enterprise security and governance, and get integration benefits all the way up to the app layer with things like the Spring Framework (for Java). When platform engineering people talk about building a platform, that’s exactly what the Tanzu Platform is: a platform.
And, if you’re working at a large organization, chance are you already have the Tanzu Platform in place. This also means you probably have generative AI in a box available to add. No need to ponder building a new stack from scratch.
Unlike a lot of the work being done right now in platform engineering, the Tanzu Platform has existed for many, many years, is used in numerous production environments with uncountable applications. That is, it’s proven, comprehensible, usable, integrated, and available. If you want to know what platform engineering is in product/stack and in practice, look at the Tanzu Platform and Tanzu Data Services.
(I didn’t, like, run that by anyone at work this Saturday morning. It’s how I think of it. So, like: opinions are my own not my company’s, etc., etc. You can read the fully corporate-pondered overview on the weblog.)
Speaking of the Tanzu Platform, I talked about it in the first segment of this week’s Software Defined Talk episode. The episode opens with me commenting on the new, unnamed expensing software I get to use. Then after a brief Tanzu Platform discussion, we spend most of the time talking about Microsoft Build, and a little bit about the prospect of Raspberry Pi IPO’ing. Listen it, or watch the unedited video. Or both!
I read some contemporary, public-facing philosophy recently. My degree is in philosophy (and English), so why not? My thought: who is this written for?1 It was just a bunch of circumlocutious writing with endless references to Continental philosophers - even Kant! And don’t get me started on how much the word “Hegelian” came up.
One could say my mind has been warped by corporate communication (PowerPoint, industry analyst reports, Mento memo’ing) and blogging…but, really: just the medium of how we communicate now - clear, concise, to the point - is a big challenge for philosophy writing, and I assume humanities writing in general.
There’s no need for that baroque style we call “academic”! You could just come out and say what you’re thinking.
But, the art of philosophy writing has always been obtuse. To make it visual, humanities academic writing is like abstract art: you have work at it to see what’s there, appreciate what’s going, get some utility out of it.
However, I think that’s one of the major points of academic writing. There’s a sport to reading it and liking it.
So much of philosophy writing - and literary academic writing - is entertainment for those writing it and reading it. Is there a dichotomy there: writing for academic perfection and pleasure versus writing for clarity and action?
If Kant or Hegel were bloggers, or, better, Tweeters, what they were saying would be so much different - the medium would change it. And, I probably would have read more of them. You could look at different mediums in philosophy to see how the medium changed the outcomes: Plato with “dialogs” (really, just very boring plays, contemporaneously they’d be movies/TV, or Broadway musicals), Nietzsche with aphorisms (contemporaneously, greeting cards, smug coffee mugs, and bumper-stickers).
Anyhow. Back to inverted pyramid mode.
I have access to the MacOS ChatGPT thing now. It's the same as the other desktop GPT overlays I've used like Jordi Bruin's stuff. Too bad for Jordi! Or, maybe, too great, cause now he’ll do some new, great features.
I didn’t use his stuff because it was just a wrapper for the web ChatGPT Plus, or you had to put your own OpenAI token in and get metered usage. I’ll see if I end up using the desktop ChatGPT from ChatGPT more. That kind of nudge to change behavior - it’s the official one rather a third party - would be an interesting…uh?…”UI bias”
I think the only "new" thing it gets you is the voice interaction which was only on the iOS app previously. I like this on the mobile app, but connectivity made it hard to use. Maybe with my wired, gigabit connection on my desktop it’d be better.2
It still doesn't have the ability to download the transcript for a single chat session. This seems bonkers.
Also, the UI is exactly like the web UI in layout, functionality, etc.
OK, there is a slightly new thing: collapsing a screenshot workflow into an app feature. Once you give it access, you can have it auto-upload a screenshot:
I didn’t give it access to record my screen. I mean, obviously they’re on their own ethical vibe over there.
All jargon is meaningful, just maybe not to you.
Coté’s calendar style: quick to decline, slow to accept.
It’d be fun, if when I fly from a distant time zone to another (ATL to AMS), when I opened my laptop it had a moment of geo-confusion just like me. Error: where am I?
“Sort of a lunch-pail occupation.” Gruber.
The obligation to argue accepted truth, as opposed to the obligation to plan your family’s annual vacation.
“If a developer hasn’t gulped down their AI pills and pulled on their Copilot pants, Microsoft doesn’t seem that interested.” Here.
Talks I’m giving, places I’ll be, and other plans.
SpringOne Tour London, June 5th. DATEV Software Craft Community online, June 6th, speaking. DevOpsDays Amsterdam, June 20th, speaking. NDC Oslo, speaking, June 12th. SpringOne/VMware Explore US, August 26–29, 2024. SREday London 2024, September 19th to 20th.
Discounts. SREDay London (Sep 19th to 20th) when you 20% off with the code SRE20DAY. And, if you register for SpringOne/VMware Explore before June 11th, you’ll get $400 off.
I’m back in Amsterdam after a week in Atlanta. The new, “expat goes back to the States” thing I noticed is that some Americans will wear a short-sleeved golf shirt with slacks and figure that’s good enough work attire. Also, they tuck in the golf shirt.
I can’t imagine Europeans doing this, or, really, wearing golf shirts at all outside of…golf. Even a t-shirt would be more “normal”: there’s something about the plastic sheen of a golf shirt and those two or so little buttons on-top that makes it so much different than a t-shirt. Now, a polo shirt…this might be acceptable if it was properly branded with a classic polo brand or “Boss.”3 I’m not sure though, I’ll have to start noticing.
And, as ever, what they do over in UKI might be more Yankee-Continental than Continental-Continental.
This feels like a fun thing to start asking AIs when they summarize text.
Also, while the reaction to the Sky voice isn’t really, like, shocking, that it took this long is weird. I’ve been using Sky since that feature came out, and I think I even commented on the podcast that it was clearly a rip-off. I guess a lot of people didn’t subscribe to and use ChatGPT Plus? Also, I miss that voice - it was great! Hopefully they can figure out getting it back. The other voices are way too…artificially cheery.
The popularity of the Hugo Boss brand in Europe is an ongoing curiosity to me. Also: Gant. ¯\_(ツ)_/¯
Just wastebook and links this episode.
If you’re a writing type suffering from imposter’s syndrome, spend some time as an editor. You’ll soon discover that writing is incredibly hard, so if you can pull it off, you’re the real deal.
“Preview The Magnitude Of Our 2024 Agenda” - subject line of a Forrester email.
Alex: “So you want training data for my replacement?” Son: “Well, at least until they unfreeze your brain.” Here.
“‘Culture’ is a kind of catch-all in social affairs that can be deployed to explain almost everything everywhere, and so it usually explains nothing at all. If the advice your theory yields is “change the culture”, then you have no meaningful advice.” Here.
“A Man Was Ordered to Build a Fence to Hide His Boat. He Asked an Artist to Paint the Boat on the Fence” Here.
“business websites with personality” Here.
“While developer led motions are awesome, all roads eventually lead to steak 🥩 dinnahs and hard-core enterprise sales to pay the bills!” Here.
“Then he gets up from the table and shrimp fall out of his pockets and he walks out of the boardroom trailing shrimp everywhere, this is what corporate finance is all about.” $11m on shrimp.
“the classic difficult-childhood-because-his-parents-made-it-too-easy” Here.
‘Yet again, I kinda sort of understand what he’s saying, but I have no idea what any of that means in concrete reality beyond, “we will collect lots of data on inhabitants, and make adjustments based on patterns.” And again, I’m not sure you need a cyberpunk city to do that.’ Here, from 21,000 words on Saudi Arabia.
“Who the fuck wanted to spend 44 billion dollars on a toilet?” Burn Book.
We need more life advice books by people who don’t have time to write one.
“manufactured serendipity” Here.
"Be an active bystander."
What Factors Explain the Nature of Software - “The only way to know exactly what software we want to build is to fully specify it: without doing so, there will be gaps between our vague ideas and harsh reality. However, a complete, abstract specification is, in general, at least as much work as creating the software itself – in many cases, it is substantially more work.”
Vercel Momentum - “Vercel was started in November of 2015 so it has taken over 9 years to cross $100M ARR - one has to be patient in dev tools and it doesn’t always work out!!! Harness started in 2017 so it too took 7 years to cross $100M ARR, but it also had a more focused model to sell to larger enterprises as a devops infra provider from the very beginning.”
BP AI use for developer productivity - “A passing mention in BP’s latest quarter: 'Due to GenAI and in-house built developer productivity tools, we have increased the output of our software developers by around 70% year-over-year” and cut external developers by 60%.'"
The Forgotten War on Beepers - Back when pagers were ruining the youth.
Enterprises struggle to show the value of AI projects - “’Organizations who are struggling to derive business value from AI can learn from mature AI organizations,' Leinar Ramos, senior director analyst at Gartner, said in a statement. But it’s a small group: only 9% of respondents in Gartner’s survey are characterized as AI-mature.”
Gartner Forecasts Worldwide Public Cloud End-User Spending to Surpass $675 Billion in 2024 - Cloud: lots of money.
Crocodilia: How Crocs became a clog-selling profit machine - I need a pair!
As clicks dry up for news sites, could Apple News be a lifeline? - Wow! // “the Beast on track to make between $3–4 million in revenue this year from Apple News alone — more than its own standalone subscription program, and without much additional cost.”
The #1 Platform Engineering Problem You’ve Never Heard Of: Platform Decay (Webinar Recap) - Avoiding the accidental platform and platform sprawl. Gartner Identifies the Top Five Strategic Technology Trends in Software Engineering for 2024 - (1) Software Engineering Intelligence, (2) AI-Augmented Development, (3) Green Software Engineering, (4) Platform Engineering, (5) Cloud Development Environments. “Gartner predicts that by 2026, 80% of large software engineering organizations will establish platform engineering teams, up from 45% in 2022.” // What’s weird about this prediction is the question: “uh so what have the PaaS, DevOps, and SRE people been doing this whole time?”
Using AI to generate web forms from PDFs - With Enterprise AI, the app is still the hard part: “What took longer was the other stuff: identifying the problem, designing and building the UI, setting up the templating, routes and data architecture. It reminded me that, in order to capitalise on the potential of AI technologies, we need to really invest in the other stuff too, especially data infrastructure.” And: “What really slows transformation is bureaucracy. It’s getting permission to use a tool like this, and to make improvements to the underlying service.”
Internet use statistically associated with higher wellbeing, finds new global Oxford study - “The study analysed data from two million individuals aged 15 to 99 in 168 countries, including Latin America, Asia, and Africa and found internet access and use was consistently associated with positive wellbeing.” // At this point, who knows? Like everything else in life, it’s probably how you use it that matters.
Why We Hate Working for Big Companies - Capitalism knows the central planning committee is bad, so, logically, Capitalism runs on corporations that are central planning committees.
Everyone hates Workday - “Workday reveals what’s important to the people who run Fortune 500 companies: easily and conveniently distributing busy work across large workforces. This is done with the arbitrary and perfunctory performance of work tasks (like excessive reviews) and with the throttling of momentum by making finance and HR tasks difficult. If your expenses and reimbursements are difficult to file, that’s OK, because the people above you don’t actually care if you get reimbursed. If it takes applicants 128% longer to apply, the people who implemented Workday don’t really care. Throttling applicants is perhaps not intentional, but it’s good for the company.” // This gets close to something, but doesn’t have enough empathy (understanding?) of executive life to sound right. More like: the purpose of a large organization is to perpetuate itself as is. It only changes if there is a crisis or a huge improvement to be had (10x? 20x?). Workday is good enough: all the effort you would put into fixing it would just result in the same general outcomes. You would take on the risk of it failing, and get little improvement of things went well.
Talks I’m giving, places I’ll be, and other plans.
Atlanta Executive Dinner on Enterprise Software Dev, etc., May 22nd. SpringOne Tour London, June 5th. DevOpsDays Amsterdam, June 20th, speaking. NDC Oslo, speaking, June 12th. SpringOne/VMware Explore US, August 26–29, 2024. SREday London 2024, September 19th to 20th.
Discounts. SREDay London (Sep 19th to 20th) when you 20% off with the code SRE20DAY. And, if you register for SpringOne/VMware Explore before June 11th, you’ll get $400 off.
I’m in Atlanta this week for two events (see the dinner one above). I’m in a Residence Inn this week - which has a kitchen (no oven, though). Usually I’d go to Whole Foods to eat from the hot bar for at least one meal, but this morning it wasn’t open yet. I realized I could just buy some eggs and salsa, and an avocado. I always buy avocados in the States: they are not good in Europe. I also realized I could buy a steak! (Hard to get American style steaks in Europe). Amazing!
//
I had a long conversation today about a gap in the enterprise software journalism and analysis world. There doesn’t seem to be enough reporting there - the focus is on crushing iPads more, AI now-a-days, sure. The problem with the space is, how do you make money publishing about it? A lot of companies end up getting sponsored by vendors, which is totally fine.
//
I also recorded an interview with JT Perry about Enterprise AI in healthcare. It was great! It should be out later this week.
//
On the way over here, some Americans complained about how hot it is. First, in Schiphol, then in the Atlanta airport SkyTram. What they meant was the AC wasn’t cool enough. Long ago, when I was traveling for RedMonk with James, he and the other Europeans would always complain about how cold it was in American conference rooms. They seemed insane. Who wouldn’t want cold AC? But now, living five years in Europe, I get it. I didn’t even feel like it was warm. And when I walked into my hotel room, it was freezing! I turned it up from 60 degrees to 70, and that seemed right. I still like cold AC, regardless of tempature. The crispness is impossible to get otherwise.
//
That Gartner prediction about platform engineering above (“Gartner predicts that by 2026, 80% of large software engineering organizations will establish platform engineering teams, up from 45% in 2022.”) has really “chipped my chops” as one of my friends like to say. There’s something weird with the current conversation around platform engineering. I get all the newsletters on the topic, read all the stuff. As I mentioned above, what have we been doing this whole time with “platforms”? Did they not exist?
Here, we have Paula and Colin talking about platform as a product in 2019. The book Team Topologies was out in 2019. A year before that, 2018, you had the Bottcher/Thoughtworks definition of platform. Cloud Foundry was released around 2011, and was at the beginning of its stride in 2015. There’s Heroku, of course, the late 2000s. And then DevOps, SRE, and centralized tools groups of Netflix and Google fame spread through-out there. I guess that’s five years since 2019. Of course, those ideas were germinating for 2 or three years ahead of that (just proposing a book, getting a book accepted, writing a book, and then waiting for it to come out probably took a couple years, 18 months at least?).
It seems weird. Lots of people don’t even automate their build. One surveys says that only 35% of respondents apps are are using DevOps for management. So maybe it is actually something new? Rather, maybe actually doing it is something new.
I look at these years of conversation about the platform topic and I get a little queasy: are we just going to keep talking about this over and over, giving it new names here and there? Are enterprises giving themselves the chance to see their “transformations” through to fruition? It feels like introducing a platform into a large organization, in a big way, takes three years, maybe even five.
I’m not sure: but there’s a good conference talk in there! Maybe I can work it out for my upcoming “Developer Productivity Is Waste” talk (NDC Oslo and then DevOpsDays Amsterdam). The premise would be something like “developer productivity is a symptom, a signal of actual problems. It’s just how you measure something, not an ends to itself.” I mean, I guess that’s the problem with any metric: when you think improving it is the point. The point is actually “why aren’t we doing better, and how would we fix it.”
In the case of all this platform stuff, based on how “new” platform engineering is, and the other surveys I’ve looked at recently (low CI/CD usage, low DevOps usage), the answer seems like: because you’re not doing much to improve it.
Really up-lifting stuff!
Internet use statistically associated with higher wellbeing, finds new global Oxford study - “The study analysed data from two million individuals aged 15 to 99 in 168 countries, including Latin America, Asia, and Africa and found internet access and use was consistently associated with positive wellbeing.” // At this point, who knows? Like everything else in life, it’s probably how you use it that matters.