How'd you like to listen to dot net Rocks with no ads? Easy? Become a patron for just five dollars a month. You get access to a private RSS feed where all the shows have no ads. Twenty dollars a month. We'll get you that and a special dot net Rocks patron mug. Sign up now at Patreon dot dot NetRocks dot com. Hey, welcome to the podcast. It's Carl Franklin and Richard Campbell and Maddie Montaguilla. It's dot net rocks. It's up, Richard, Maddie.
Good to be back home.
Galavander, Sir Galivant.
Spent a couple of weeks in Mexico and Acapoco, you know, soaking up some sun because it's awful gray in the in the Pacific northwest, right, and then stopped in Mexico City to visit with Paul Throt and his wife. A few days in Mexico City, and then we couldn't get out. That's cool, because the Aero Rexico kept canceling our flights over and over again, like two days running. I finally just bought a ticket on Air Canada and did the
long way back to Vancouver through Toronto oh god. But the issue was that it was foggy at YVR and for what ever reason, Aero Mexico I guess they just won't fly in the fog and the air Canada guys are like, let's go. So they got home that way. But you know, I figured it out.
All right. Let's start with what happened in nineteen eighty seven. This is episode nineteen hundred and eighty seven. By the way, we're moving closer to two thousand, which you probably know we're going to do live at the party with Palermo. Yeah, MVPVP seven And we've asked you for your little memories. I guess, yeah, stories, stories, Okay, okay. Nineteen eighty seven, Wow, A bunch of things happened.
What are your favorites?
Well, Black Monday happened, stock market.
Crash, right, a big market correction.
Dow Jones industrial average up twenty two point six percent in a single day. Of the Princess Bride, Dirty Dancing. Aretha Franklin became the first woman inducted into the Hall of Fame. You two, Joshua Tree, Right, all good stuff, easy music, yep, yep, yep. Let's see. British Airwaves was privatized and listed on the London Stock Exchange. Ronald Reagan and Mikhail Gorbachev meet to discuss nuclear disarmament.
All right, let's start in space, of course. We last week's episode nineteen eighty six was the Challenger disaster, which grounded all shuttles obviously, and began.
The geez, you guys have a tough few years here, huh.
Yeah, it was rough.
Yeah, although obviously there's no spacial flights in nineteen eighty seven, this is when the Presidential Commission on the accident is going on, they called the Rogers Commission, and the reports and implementation are put out in nineteen eighty seven. It starts with the solid rocket mooter redesign three to zero rings instead of two, and a heater system to make
sure that those seals never get so damaged. You know, they challenge your incident was the most disastrous of the problems with the solid rocketbuster, but it wasn't the only one, and that technical solution worked, there would never be a problem with the boosters.
Again.
The harder one was the cultural changes. So they were supposed to have an independent safety organization, which they definitely laid out on paper just didn't actually ever use it. And then this concept that the Rodgers Cousoul talked about of normalization of deviantce this sort of recognition that just because you didn't have a disaster involving a problem doesn't mean it's still not a problem. It should be immediately addressed.
So engineers supposed to be able to speak freely that whenever you have an anomally you take it as a warning. This obviously doesn't work, because we'll have the two thousand and three Columbia disaster where they had had lots of flights where foam was coming off the external tank and eventually punches a hole through the wing of Columbia and dooms the vehicle. They also add the crew escape system, so this is an escape This puts the shuttle astronauts
into proper pressure suits. Up until then they've been just in flight suits, and also adds this escape pole out of the hatch. In theory, during a disaster, you were supposed to clear down to the lower deck where the door was, blow the door out, this rod would come out, they would hook to it. That raw would help fling them beyond the reach of the shuttle so that they could parachute to the surface.
Funny thing about explosions, though, you don't really have time to do.
That, not a lot of time for any of that. Plus most of the flight profile of the ascent does not allow this to work. So even though they've added all this complexity, most of the ascent was still non survivable.
So you think it was more like, hey, we can spend money on safety watch.
This and one of the things you see overall and the results of the Rogers Comission, we have all of this in hindsight is mechanical they could do, they did, and the cultural changes just they couldn't pull off. But arguably the most important thing that comes from the Rogers Commission is acknowledgement that the Special is not an airliner. It should not be operated about an airliner and an
experimental vehicle. The flight rate should be much lower, lots more testing done, which just raises the cost of all the flights. And of course they would follow that with much lower flight rates for a few years and then crank them back up again as the demands for the International Space Station so come into play. It's great that and I went back and read this again. There's a whole couple of paragraphs to talk about. Look, this was not a technological failure. This was a cultural failure. And
the only real solution is to change the culture. And that's the hardest thing to do. It's right in the dock, and it's exactly what.
They couldn't do ultimately.
Other than that, Like, it's a pretty you think it was a pretty slow space here, right, the Americans only through three They had one flight of the last flight of a Titan three B lifting an Air Force communication satellite. There are two dental delta flights, one for a geostationary satellite, one for a communication satellite. But then there's the Soviets. It's nineteen eighty seven. It's kind of the height of
the Space Defense Initiative SDI aka star Wars. This is something that Reagan announced in nineteen eighty three and was formed in nineteen eighty four. It'll become the Missile Defense Agency in another decade or so. It will never actually
fly any hardware. In fact, in eighty seven, the American Physical Society concludes that the technologies that they're talking about, which was flying all these little satellites that would track and shoot down ICBMs, is decades away from being true, but the Soviets don't know that, and they attempt to test the defense system. This is the Energy A Rocket Booster. This was the first flight of the Energy A Rocket Booster. Obstensibly it was built for Barran, and Barran was the
Soviet space shuttle. It'll only fly once in eighty seven, but its first flight. They because it can lift ninety five metric tons like that's an incredible payload that basically can't be matched at Saturn five level payloads. The most powerful rocket the Soviets ever built. They lifted this vehicle called Pollius.
It was supposed to just be a.
Test payload, but it wasn't. What it was was a weapons platform carrying a megawat carbon dioxide laser, but it never made it to orbit. It sounds spooky. It's dangerous, is what it is, and the failure to get to
orbit is weird. The Energy of booster worked flawlessly, but the vehicle was actually mounted in the launch platform backwards deliberately, so as soon as it's separated from the booster, it was supposed to turn around one hundred eighty degrees and then boost itself to the rest away in the orbit and instead of turning one hundred and eighty degrees, it
turned three hundred and sixty degrees and deorbited itself. The debate is whether or not it was an accident, because the argument here is as much as the Americans were talking about defensive weapons systems in space, they had flown nothing. This was the first weapons platform ever deployed in the space, again in contravention of the Outer Space Treaty of nineteen
sixty six. And so there's this whole subtext that maybe when Gorbachev figured out what they were doing, insisted that they make sure it doesn't get to orbit, and such a simple thing to say, instead of turning one hundred and eighty turned three sixty, so that everything worked perfectly, but it never flew. Now, a lot of the technology was developing. The best argument I say in favor of this is the fact that they didn't just build one test platform. They had built a bunch of parts, all
the control systems, like a whole lot of them. The laser actually came from an anti ballistic missile system was supposed to fly on an aircraft, and those parts continued to be used. The Mirror Space station, which had gone up in nineteen eighty six, and I think I may have failed to mention that a bunch of the additional modules that went onto that all used parts from the
Polias project. In fact, even the two of the modules are on the International Space Station their technologies directly derived from this project. So this project was not experimental. It was a large scale, long term project that got repurposed away from being a weapon system.
You know, I forget about writing the history of dot net. You got to write a book about this and sell the screenplay because it's just like, I'm like on the edge of this is.
The craziest thing that ever happened that virtually nobody talks about.
Yeah, I've never heard about it.
Can you do it?
It is a full episode of this. Like, I don't need the text any of it. I'll just sit and listen.
Yeah. Well, we do space geek Out every year at the end of the year, but this wasn't on it.
I mean, it's mostly about current stuff, not this historical stuff. But Polyia stands by itself. It's insane what actually happened there. Well, also, think about a Megawak class laser in nineteen eighty seven. Yeah, like honestly star Trek stuff on the other tech side of things. So this is the year that the GSM standard gets formalized as the European Staminay for telephanites where
SMS comes from, text messages comes from. Of course, the American standard the Quenee Center was the CDMA, which will ultimately lose out to GSM worldwide. Of course it won't reach this is only the standards point eight eighty seven they producle. We're still market till the early nineteen nineties. Computer wise, this is the year of the IBM PS two.
Oh boy.
This was a three eighty six megabyte machine. But the important part of it is that it came standard with three and a half inch floppies and the Vega video standard videographics are right, really setting the standard for both those. This was considered my IBM's push to try and get control of the PS of the PC market that they invented in lost large control. They sold more than a million of these, like they did well when.
I had my first Vega card and monitor. The most I was just thinking of this yesterday. The most amazing thing was seeing a depiction or a scan of an actual photograph that was beautiful. Yeah, and you know, like a woods scene in winter or something like that, and I was thinking, oh my god, we've come so far.
Yeah, the color palette was so much larger. That was like a big deal. It was a big deal. So this is also the machine that you had a choice between Windows two, which was released that year, and two STAS four or O.
S two version one.
And Windows two was you know Windows when had been out, like it was still pretty primitive. Guey, So OS two is a reasonable option that So it's all the beginning of that. Uh yeah, I got three more for you. Language wise, this is the year of Pearl. So Larry Wall ships Pearl. He originally built it to do text file parsing and report generation, but we know what it really became. And this is the quoted term the duct tape of the Internet.
Well and also it reads like a cartoon character swear.
Yeah, so it was it's you know, it's it's just it's a text parsing tool, right, which is perfect for the Internet. But yeah, Pearl kept a lot of Internet working, a lot of websites functional, Like Larry Wall did us all the service, but it's a very challenging. H here's why Carl's going to love. This is the year that the company called ad Lib from Quebec City is formed to make the music synthesizer car, typically known as the car. Right they will be bankrupt in five years.
Yeah, Creative Labs flete at their lunch.
Creative Labs will clauter them. But they prove the market or high quality PC back then.
You know.
Now all motherboards come with audio, but back then they did not. Right, you had to get an add on card.
One last one out of compu serve by a guy named still will hedwork is Steve well Hide working for the company creates a graphic graphics interchange format or JIFF, not giff if not Yiff, because Steve himself the guy it said, choosy developers choose.
Jeff come ondraschange format graphics.
I'm just telling you what they created. But by the way, guy died to COVID nineteen. That's a shame. But he didn't he.
Sorry, wanted everyone to use it.
That's wheet and he was he bid to play on choosey kids, choose Jiff, Yeah, choose Yeah, Yeah, that's choosy mom, Jeff, Yeah, all right, cheesy mount Jef says he used that same line to get you to use this graphic format that would where it's a it's a non resolution losing a non destructive compression format like it was a great it's still great. Anyway, there's my stories from nineteen eighty seven.
Okay, now that we've wasted thirteen and a half minutes of your time, it was no waste.
No, I'm kidding, Sory, I talk too much about Poulias. About my god, what a coolnes is is so cool.
That no, I'm serious that that's the best story I've heard this year. And you know we're just getting started.
All right, so history of space geek outs let us know. Maybe I'll write a couple all.
Right, so let's roll the music for better no framework, go all right, dude, what do you got? I got something that Jeff Fritz and I wrote. Oh nice, Yeah, Mattie's mouth just went open.
Okay, oh yeah, No one's consuming, but I have a feeling I know what this is. I'm very excited.
All right. So we basically have been experimenting with various SEQL server MCP servers right because we want to be able to ask our you know, copilot or other LLLM tell me about this database and you know, and really what it is, it's to allow a customer to ask questions about a database and get a report. It turns out that's a difficult problem and most of the MCP servers that are out there all require you to hand them a connection string, and I'm like, no, I don't
want to give you my connection string. That's now I got to put up all these guardrails and you know, do all this other stuff tell you not to do, and I know lllms don't always do what you tell them not to do, or they switch that revers it. So we came up with this idea that we want to first generate a schema from a database connection string in Jason and then create another Jason file that has all the you know, tell us about your database, these tables and what are the aliases that we can call
them the fields? Do they have aliases as well? What fields are protected? And you should never return them? And the whole idea is we want the LM to use this data to create a system prompt, and from that system prompt, it's got the schema and it's got all the other information it does not have access to your database. It will generate a select statement, not a delete, not an insert, not an update, only a select statement. Fun yeah yeah, and from that you can now you get
what you get. You got a select statement, you can execute it. So the it's great because it's safe and it's performance because you're doing all your loading up, you know, on the fly beforehand, not on the fly, you're doing it beforehand. But the problem is it takes a lot of token so you know, if tokens.
You guys are gonna have to buy and All you can Eat package To make this make sense, every developer I talk about is really use these tools. You're not only in the all you can Eat package for various services. You're getting calls from a regular basis going what are you doing?
Carol and I were talking about Sweetwater the music website before, and it's like the rep that calls you and is like you want to buy something? They need that for like you know tokens, Now, hey you need some more tokens. I saw you had a project.
I've got some tokens.
It certainly makes the case for some time in the near future we're going to be back to running local hardware to handle these things.
It's funny you should say that because I just ordered a seven thousand dollars PC gaming PC just to run Olama. And maybe I'll use the foundry when that when I can figure that one out. But the whole idea, yeah, let's let's do this local, because then tokens aren't a problem, not going to run out, and you know, now I just need a BFI graphics card and a BF computer. But you know I'm there, you go.
Yeah, I've already talking to development organizations where we're talking about putting together a couple of racks of machines, and they're calling me because I actually used to do this stuff and you can't find a lot of people who do it anymore. Saying if we were buying some HP two hundred, it's like laying out two or three racks worth of gear for a team of thirty that's all shared,
you know, within can we keep it running? Doesn't make sense, like this is actually more cost effective buying that kind of equipment than it is to pay for it. And it's like right now with the all you can eat packages at that scale, it's not, but it will be because this is all you think cannot continue, Well, the.
Other benefit, just besides tokens is keeping your data local. Right, My customers don't want me sharing the data based schema with you know, cloud based llms. So all right, well it's called AVN Data Genie and it's on my GitHub. So and also we did code it with AI episode fourteen about this and you get to see it in action and it's pretty cool.
It's interesting. Good on you guys.
Yeah, cool, that's what I got, Richard, who's talking to us today?
I got to comment off show nineteen eighty one, and that's the show we called The Role of AI and Software Development came out just before Christmas. We had Jeff with us and Bill Wilf and we did in front of an audience which has been very rare. Yes, and we had a really great group of people involved in all that, and of course listened to a ton of comments, not the least of which was ones going, boy, this sounded like it was going to be a turkey but
it was really cool and thanks for that. So it's great, you know, because it's awfully pretentious to talk that broadly, but I thought we had the right people in.
The room, Like I agree, good conversation.
For it, and David grabbed on a comment that I said, which was, lots of people going to get hurt, but I don't think it's going to be us as rather naive. In the long run, it'll all come out in the wash. But right at this moment, developers are losing their job at companies all over the world as CEOs have been promised to dream where a large provider using unspecialized developers along AI is a suitable replacement for their subject matter experts in ten years time. I figured it'll all be
figured out, but that doesn't help developers today. Work for companies have been sold a lot. You know, I'm not going to disagree with you, David, because it's not just developers. Lots of companies are using the excuse of AI to layoff workers, and arguably some of them are even doing experiments to see if this technology can actually replace workers, and for the most part are failing because these aren't replacement for workers, they are work enhancers like most automation is.
I would also point out that we saw weirdness in developer employment before chat, GPT and the llms took off. You know, the pandemic RETAVOC and it seems like the tech giants as a whole have been using the threat of layoff almost as a weapon to keep their employees scared. Yep, right, you know, and which I find ridiculously offensive. But I but it's working out for them. They all have record valuations.
Like God, I can't wait for I can't wait for the a bubble land and just for people to get sane, you know, Like you know, the upside to a recession, as a guy who went into the workforce in the eighties during a recession, is that you focus on value, and right now we really aren't. We're focused on speculation and insane levels of growth that are unsustainable and damaging and often stupid, and lots of people are being hurt.
But I made that comment. I don't think it's going to be asked what I was thinking long term in the sense that we are used to building automation childs for improving people's productivity and using automation kills for protype and so we typically get to there faster than most By your right, and I completely agree with you, a lot of people can be hurt on the way.
Something about it is just like the multidisciplinaryness of it like, I'm a PM, but I have a background in software and so, you know the role of a PM is kind of in jeopardy too, because the engineers have time to actually care about the stuff that I care about now. Yeah, and so it's really fascinating. But I do think, like you know, it will come out in the wash. I agree with that it's going to be painful, but I
think we'll be fine in the end. The engineers are problem solvers, so solve the problem, guys, you'll figure out.
Yeah. You know, the Internet shook everything up. They move to mobiles, shook everything up. The cloud shake and everything up, and there were job losses all across those you know, making run as through the migration to cloud. The number of assisted men's going. They're letting me go because they think the cloud will run it all for them. Then got hired back in one form or another, but also with their skills and mindset changed. You know, my job
isn't to rack and stack servers anymore, it's to manage them. Yeah, David, thank you so much for your comment. Obviously it's an ongoing topic. It's going to be all through shows in the next few years as this settles out. So we appreciate your feedback and a copy of music Bias on its way to you. And if you'd like a copy of Music co Buying, write a comment on the website at dot NetRocks dot com or on the facebooks. We
publish every show there. Any comment there and I read it on the show, we'll send you a copy of music Cobe.
Okay, now we can talk to Maddie and I'll introduce her. Formerly Mattie. You heard her chiming in there. Mattie Montague is a senior product manager at Microsoft and the product lead for Aspire. She's been building dev tools at Microsoft since twenty eighteen. Before Aspires, she worked across the dot net stack and brought dot net Maui to life. Maddie has been obsessed with making developers' lives easier ever since her first dev tools internship, where she realized that perfectly
combined her high tech background with her passion for product development. Officially, welcome back to dot net rocks.
Yay, thank you. I'm excited to be back.
I gotta tell you, I think the last time I was here was like three or four years ago, and it was an in person recording.
No, it wasn't three or four years ago? Was last year at a dev intersection where you and I aspirefied dot Net Rocks.
Well that was oh though, that was a Rocks episode. Yeah, right, you and David talking about Maui. Yes, maybe it's Doug.
Yeah, yeah, that's right. And we got to jam.
Yeah we did. Oh my gosh, I was crazy. I would love to do that again.
Now now that I've I've dug into your repertoire much more so, I think I can keep up with you and Chip Now.
When I knew you were moving over to Aspire, literally my reaction is, oh, I guess it's going to be a real product. Is now it's going to have adult supervision, like somebody has to rain and Fowler goodness knows. And then one thing I know about Maddie is she can. Yes, she could wrestle these guys of the crowd. They're scared to death of her.
Yes, it's I now say my favorite my favorite pastime is arguing with Fowler and Damien. It's like the highlight of every day.
You've made whole shows around exactly that.
Yeah, I mean when I when I was looking at moving because MAUI touling kind of was in a good place. I was like, I need to do the next thing, right, Like, what do I do? I don't, I can't. Dave does
so much. I did so much, Like we have these junior pms that were coming up Maui And this was like two years ago, and I had like a crisis of faith and I was like, I need something new, and so I went to my SKIP and I was like, the only thing I'm interested in working on is this Aspire thing, because they clearly have no idea what they're doing. They're building something cool, but I don't know what it is. But I don't know anything about.
The cloud, so it's probably not a good fit.
And it's kind of an asset coming in because you don't you know, there's an awful lot of the people who need Aspire who also don't know.
This right right, And so he was like, well, you know, there's a lot of strong personalities on the team, and I was like those two, Yeah, okay.
That's that's what you call strong personality. Okay, so yeah, there's the least of your concerns.
About a month in my now manager Glenn, who's you know, also been around and was part of the dot net core thing with the two of them.
But Damien Power, Yeah, he's the best, he says to me. He goes.
It turns out all we needed was to put someone from Boston on this team. He was like, it's so funny. You know, you're just you're just not You're not very You just don't deal with you don't put up with it. I was like, yeah, I'm very.
Yeah, you would be correct.
And and my experience dealing with Damian and David, I'm sure they're delighted.
Oh they're the.
Best, absolutely the best.
They're like my two big brothers. I swear like I rib them like a baby sister and they totally take it.
And I've learned so much.
But yeah, but I guess that this puts you on the spot. Then it's now it's your turn. Define Aspire to us. What is this?
Yeah, Aspire as a tool for building apps, for streamlining app dev local first all the way through to the end.
And it's type safe.
So it is a is a type safe manifest it is a CLI. It's a set of packages that you can use to build a what we say distribute an app. But that's a website with the database, right, like at the end of the.
Day, and I called it, I think I still do a visual studio template on steroids.
That's I mean, that's a pretty good way to put it. It's it's you know, it's a project runner, it's all these different things. It's it's a it's a socker composed but like times a million.
Because it's also re entrant, right, I mean, the templates always a starting point and then you have to go on from there and you're never going back, but you can still go back to the Aspire tools and tweak things and right, you're not going to destroy the work you've done.
Yeah, happy to tell you that a V and data Jenie was done with Aspire. Good. Yeah, yeah, of course Jeff worked on it.
And you two did that aspire ify dot net rocks, yes at their last fall, which is hilarious.
Yeah, we do those. We do streams on Friday. It's the shameless plug.
So Me, Fowler, Damien and the cast of characters from the team usually will pick people's apps up and try to add Aspire to them or implement a feature or whatever, and it's always a complete mess and it's very funny. And now we play bingo during it, because there are just certain things that happen every time.
Yeah, and so.
You get a card and who hits all of their talking points first or you know, yeah, no.
One of them is like Maddie swears that happened regularly.
You can put that in the center because that one's gonna have an air.
One of them is Fouler going wait pause, wait pause. So we have a good time on this.
But it's brilliant. Yeah, it's great to see the team. Hey, they're all lovely people, and like, these are the people making your product. Yeah, and you put yourselves into jeopardy every time, Like, I presume the product gets better from this practice.
Oh yeah.
We had someone ask us like, so, do you guys do it just because you like want to make the product look good?
And we were like, no, make the product look good.
It would not be this, no way more plan than that. But but we've seen this over and over again on doing various live things. It's like eating it on stage is good gets good because it's life. Yeah, we all eat it, yeah, right, and fighting through that and getting out the other side. That's how you learn to trust stuff. It's like, sure, it's not going to be easy, but we're gonna get there.
Yeah.
So Aspired debuted with dot net eight, right, and then in dot net nine more stuff, and in dot net ten, more and more stuff. So maybe we should go through some of the things that have evolved and Aspire since dot net eight.
Well, when dot net eight happened, we didn't have chat GPT.
It changed a lot, and so Aspire distributed apps, right like now AI is just a new type of distributed app endpoint or back end or whatever you want to call it.
And so Aspire like kind of accidentally.
Made a really good story for that, Like you were talking about foundry and getting that running locally, Like, it's a pain to integrate with an exist app and stack. It's good to have it running in isolation, but once you need to build it into it a system, it's way more complicated. And Aspire basically turns it into another API, like another project, right Like it totally simplifies that. The other thing is we dropped the name dot net from the title.
Just call it Aspire.
That's the big thing that's happened.
Just call it Aspire because we're the dot net team, you know, we don't pretend, but everyone writes apps in multiple languages, Like, once you get to a company of a certain size, there's no way that they standardize on one language.
It's just not reasonable.
But I mean, at the minimum you were running C sharp and sequel at the best of times like this. But the reality is there's probably some JavaScript. I bet there's some JavaScript like that's kind of unavoidable.
And with with AI pythons everywhere because people are writing you.
Know, JANKI scripts that make something do something.
No, and if you've got some badass optimizers, they've written some C plus plus some rust and step get interface in front of it so you don't have to touch it.
Yeah.
If you got some weird sunk guys that are out there with some little Haskell or some f sharp, and again it's just an interface. You off, you go, Yeah, I don't care about your language. Yeah, just wherek in which you're most effective at and give me an API.
I'm good to go.
Yeah.
And so you know, we were calling it dot Net Aspire because it was started for dot net, and then we thought, yeah, job with her front end Python whatever. Obviously I'm sure you too are fully aware because your show is literally called dot Net Rocks that a lot of people don't understand that dot net is cross platform, open source still still. So yeah, we did a user studies when I had just switched over to Aspire. One of the PM's in our team, Claudia, was running them.
She was actually trying to get us to completely rebrand and we were like, that's no, like, it's we.
Can figure this out.
But she, uh, she had someone do like an interview and she was like, what what would you think about if someone asked you if you could, like, you know, do a job where there's dot net. And they were like, well, I don't know how to write like a linked list from scratch or anything, so I could never And I was like.
Girl, what, that's crazy, that's the opposite direction. Yeah, I didn't say assembler, Yeah I know.
And so there's all these crazy misconceptions, and so these dot net teams were pulling in Aspire left on the right because it's this crazy good onboarding tool. It's this crazy good like code safe tool now with agentic dev. It gives Copilot Guardrails or clod Or or your agent dujor, and it gives it a really easy way to talk to the running all the different running pieces in their logs and everything. We have this MCP. It's just like phenomenal.
But the JavaScript teams were like, we're not touching that. Let's gross.
Yeah, and so we were like, we have to we have to kill the dot net to save the dot net in it. That's what we did, and we build the features into dot net.
Well, aspires a good name.
Yeah, it's hard to google, but it is a good name. And now we have aspired dot dev the domain, so that's the handle, so that works out, you know, put on my social media.
We're out of the long multi syllabic foundation era of naming tools and I'm glad.
Well the only reason they got dot net aspire is because they put dot net in front of it. So we had to go back to the table and be like it's such a generic term, it doesn't matter. Took a time, wow, Yeah, but it's been it's been a
crazy journey. And now we're looking at like how we can break further into those other ecosystems, including like could you have the app host itself written in typescript and it Jason our PCs to dot net as the thing that's actually orchestrating everything and kicking it off, so we don't have to rebuild as fire. But we're basically like, could we have like a typescript shim so you don't
even have to have a C sharp file? Like how much trojan horsing can we do just to get to a place to help you exactly, just to help you help yourself. You can lead a horse to water, but you can't make it's all the dot net run time, kind of.
Like we need to dig a break. Yeah, let's do that. We'll be right back after these very important messages. Hey Carl here. You probably know text Control is a powerful library for document editing and PDF generation, but did you know they're also a strong supporter of the developer community. It's part of their mission to build and support a strong developer community by being present, listening to users, and sharing knowledge at conferences across Europe and the United States.
So if you're heading to a conference soon, check if text Control will be there and stop buy to say hi. You can find their full conference calendar at dubdubdub dot textcontrol dot com and make sure you thank them for supporting dot net rocks and we're back. It's dot net rocks I'm Carl Franklin, That's Richard Campbell, Hey, and that's
Mattie Montague and we're talking to Spire. Yeah, Aspire. I asked you like about the sort of the evolution from dot net eight and into nine and ten, and you know where were the most significant updates in that evolution.
Well, we we've been moving fast.
I think the first big thing that was a huge change was the CLI.
We shipped a CLI and so we we were like.
You know, why do I have to why can't we have a tool like everybod else everybody has a CLI tool.
That'd be fun.
Now you can just Aspire run instead of dot Net run, appost project or whatever it is. And that was the first kind of custom Aspire experience for tooling we built that wasn't just like the VS stuff and the extensions and whatever that was that was crazy.
We did not know what we were getting into, but we got into it. That was probably the first big.
Shift from the original Aspire, right, Like I think everything that shipped originally was crazy, and the iterations on it have been insane, Like the amount that just the integrations, like the packages that have our opinions in them. The a month that those have grown and matured is a lot. It's been a super incremental.
I mean, it's interesting to have that development angle of it's opinionated in the sense of thinking distributed specifically, you know, ideally for cloud right, that sort of cloud native mindset where you're pulling in services and so forth, and so it kind of gives you a set of scaffolds to go down the right path on that, yeah, which is very language and platform agnostic actually, but it is still opinionated, right.
And the funny part now is that and then AI stuff shows up and you know what it needs.
Opinions, opinions guardrails.
Yeah, yeah, like you're just trying a right place right time. If I'm sticking an LM in the playpen of Aspire, it likely not hurt me less, Like it needs those opinions to kind of do the right thing. I'm trying to get to a nice cloud native design, thinks yep. And here's a set of guardrails, yep.
And it has a hard time managing like multiple terminal processes. So it's like instead of NPM run and Docker and the boat, it just Aspire runs it. It's one command, right, and so the CLI accidentally really helped us with that too. And then the next big thing we shipped a version of co.
Pilot, and that's the az Dev cl I.
No, the Aspire Cli, the Aspire Cli, Aspire if you will Aspire run. We have shirts that say it.
I got two new Englanders on the show at the same time.
I get all this accentrific for you, wicked piss.
And then but a year ago we started working on co pilot in the dashboard because we were like, well, everyone has a chat now that you ask questions. MCP hadn't really become I don't know if you saw this the other day, Miguel posted MCP as the new soap, right, that's unfortunately, very good tweet Miguel, He's right, but it
wasn't a real thing yet. So we were like James Newton king on our team, like went away from the holidays and came back and was like, so I built a co pilot in the dashboard and we were like what. So we shipped that and by the time that was out the door, MCP had started to catch on. And now we have an Aspire MCP and so that will you know, can to the dashboard and it can talk to all of your resources at once from whatever your
coding agent environment is. And those have been the really I think to me, those have been the really big moments of like things that we've shipped that have completely changed what Aspire was since we started.
There's a zillion incremental things, and there's a lot.
Of under the hood stuff in the app post that changed, and how we actually run things. We overhauled our JavaScript and Python support so they actually felt like more natural. That was a really fun exercise in untraining ourselves from all of the things we thought are true, and also writing casing completely differently because JavaScript is the camel case.
Yeah.
Nice, Yeah, But I got a question, Okay, why thirteen?
Because why not?
I could think of a bunch of why not? Now what exactly you're talking about?
Like you put out dot Net ten and so you figured you have Aspire ten and said we get Aspire thirteen.
No, No, because all you people, and I'm saying this with.
Love, who do you call you people?
You people, the people of the dot net community. All you people kept going, well, I can't use as fire or nine because I'm on dot net eight. And we were like, it doesn't matter. We just sinked up versions because it was easier, and everyone was like, well, when as Fire ten comes out, I can't upgrade to dot net ten, or like dot Net nine is in an LTS release, so I can't use a fire And we're like, doesn't matter, it doesn't matter. So we said we're going
to change the version. I wanted to go back to one. I thought that would be cool instead of dot Net Aspire nine. It'd be a spy.
So you figured you wouldn't get any emails saying I can't wait till dot Net thirteen.
Yeah.
Well that and we were sitting in a meeting and I was talking about it, and they were like, you can't on you can't downgrade, and you get package even if it's a newer urion, right, so we'd have to break all of our packages. Yeah, you have to get So we were like, well we had to go higher than ten. Eleven feels like we're just kind of like, WHOA, I don't want.
To be ten? Yeah, and why would you pick twelve when thirteen is right there? Like that's way more fun.
This is the reasoning, this is the actual reasoning. We were sitting I love it, and I was in a room actually in Seattle. I was out there and we had Mitch, who's in Australia, on the phone and he goes, why don't we just do thirteen? And me and him like lock, I aswer the team's call, and I'm like this is terrible.
Yeah and brilliant. Screw them.
I remember coming home and my husband was like, why did they let you make decisions?
Like you should not go out decisions? And I was like, it's great.
And then we you know, we made stickers with little black cats on them, and we're like super into it. We're going to do something for Friday the thirteenth, and there's two of this year, there's February and March.
Right, we don't know what we're gonna do yet, we're gonna do something.
At the next meeting where you're thinking about a name for something, bring up Duncan.
Yeah, yeah, I like that.
The real question now is like what's the next version, Like do we go to fourteen or do we stick with.
Only be primes? We could go to seventeen.
Yeah, no, rules, that's it.
I love it. Only primes that's great.
But I to your point, this is the version we adumpt the dot net name where you focus on things like Python and JavaScript, being first cross members like you're being the Aspire you want to be and tying yourself to version numbers is bad. It's just really bad. Yeah.
Were there any significant bug fixes in the last version. Yeah, okay, we have.
We're I've learned to be proud of our bug debt because that moves that we're moving fast, which you're kind of have toy right now. There will be a contraction phase where everybody goes and makes quality the most important thing again. And we've done that kind of periodically over the last year, like we've taken a release or two
to just be like bug fixes only. There were a lot of weird things with SERTs and getting open telemetry wired up, especially cross language, and Damien and David Negstad and Carol on our developer control Plane team, they also have spent a ton of work basic like figuring out how to optimize devserts for all different languages, container orized or not without being too heavy handed on your machine.
And it was I have no idea, I have no idea how they made it work, but they made it work, and so it's kind of a feature.
But it felt like a bug fix.
Because you would run something that you'd think would work and it didn't.
Yeah, but it's very much how you actually serve developers as you go where they are, and the sart strategy that a Python developer may be using in his organization. If you're not supporting it, that guy's out. Yeah, you're gone. You can't interfere with my right plumbing, right, the homework of whatevery, you know, because he barely understood when we got set up in the first place. Nobody understands. That's just reality. But if you don't work with their sert strategy, they're out.
It has been fun to see how spoiled we are as dot net developers, Like we all love to complain, you know, we're developers.
We complain.
But you have dot net devserts and you just run the thing and it just does the thing right. Nobody else has that wow, or there's you know, eighteen million branches of how it works, and something like typescript or gihon a script it's like, oh, you could do it this way, or you could do this way, or you could do it this way, or you could just set this environment variable. That just makes it so that no TLS is authorized anymore on your machine from the node run time, And it's like right.
Uh oh, well yeah, and everybody, everybody has a different way of approaching this, and a lot of organizations and a lot of different techniques. They literally punt it's like use whatever you want, which another way of saying welcome to hell. Yep, yeah, right, no, two ways.
And the fact that I won't be responsible.
Yeah, well Microsoft did the right. The thing about dot net dev SERTs is brilliant free free, right, you know what you do to not have to ask the boss for anything this, yes, and so that's what they'll do.
Yeah, and you can still do it. It's using under the hood.
It uses the things that you use, right, Like we're not reinventing the wheel, which is a problem.
We've definitely gotten into in dot net in the past.
I think like where we'll be like we can do it better, like we're going to build a whole new thing for this, but we've we've gotten far away from that, especially like you know, in my decade or so a little bit less being in dot netland, where we do we do it so that you can always break out and do what the thing is under the hood if you want to and you don't want to use our stuff, and we carry that opinion into a spire where we have these opinions and these packages.
But if you want to just write code to do whatever, we don't care. Go nuts, have fun.
Nice what's next?
Huh?
And what's left to do? Oh my god?
Well you're thirteen point one. So and by the way, Mady, I just get a thrill every time I hear the PM version of you come out with is, Hey, we're going to do fixed versions, and we're going to do feature versions, like you've just been down this road so many times, Like I hear those sounds. So thirteen point one is out. You big up to thirteen. We use a heavy left. That's a whole lot of features. I got to think that the one is okay? How those
actually land? Like what did we get in? Like this is it's a stabilization rev mm hmm yep.
Thirteen one was a lot of okay, Like these are things that we because we wanted to ship at dot net cof right, because even though we were like, oh, our whole thing is we're dropping dot Net. Like writing the keynote script was very weird. Because I was like, it has to be like we're dot net and we love this and like this is not bad, but also
we don't want to lie. Yeah, and so we you know, we totally sank that up and then we shipped thirteen one like three or four weeks later because we just there were things that we were like, yeah, you know that didn't get in that could be cleaned up. We're also trying to react as Foundry has been developing like crazy. We want to make sure we have good experiences for that. And so our next release, thirteen to two is coming out.
Did it What month is it? It's January? Yeah, I don't know what's come out, but.
It's January at my house.
It's not out yet.
Yeah, so time is illusion lunchtime, doubly sing.
That's true, It's true.
That's definitely more of a stabilization and a docks release, Like we're really hunkering down on it because we build our own website. So now we have docks on our own website, right, so we're making that easier. We're looking at, you know, how to get the team to be able to contribute docs more without having to have our one four docks guide just right at all? Right, and then and using AI for that of course?
Is that is that a learn engine? Aneath at it?
We built our.
Own wow wow, Yeah, we we had a lot of arguments about it. Part of it was just because we wanted to build our website without dot Net and deploy it with a spire using the internal deployment stuff, which is a nightmare.
This is you dog fooding.
Yeah, and we wanted to be able to say we have this aspire ified thing that isn't dot net, right yeah, and learn, I mean learning a whole bunch of other stuff.
But yeah, you're not going to lift a whole learn engine into that, So we kind of got to do it your way right.
Control the brand, Yeah, and we thought it would be fun.
Yeah, the brand control is great, And like we're using Astro and Starlight, which is really cool. They actually just got bought like a week ago, but now we're like one of their showcase sites and so it's been a really cool way to like build out networks outside of our usual realm and it's really easy.
It was super easy.
So but anyways, doc debt a link to that to Astro build Like folks need to know.
This is cool. Stuff.
Yeah, and it's open source by the way, Microsoft Slash is fired dot dev on GitHub. So doc stet is this release. And then we have to get back into two. There's two big things. One is deployment. We have a lot of we've we've kind of settled on a direction, but it's not done. Aspire has always been local first, and then we've made it using AZD or whatever like
pretty trivial to spin up a dev environment. But obviously most people want to put things in pipelines or hand them off to their DevOps team, or run checks against things and CI and et cetera. And so we've started this concept of as fire pipelines and build steps and all these things, and we need to polish that up and actually finish it and build out stories outside of Azure. Because we are multi cloud, you can use AWS and Vanilla, Docker and Digital Ocean with a spire.
But so GCP will work fine then too.
Yeah, yeah, exactly, and it does work now, it's just you have to spit out the app posts as a manifest and then translate that into whatever the artifacts are. And so that's a big thing we're building out. The other thing is obviously AI. Both using and building AI, and so the MCP is crazy.
A Spire MCP and the.
Playwright MCP together, My god, it is truly like hands off because Playwright navigator.
I am so in danger when I play with the Playwright MCP of losing a day every single time I touch it. It's an instant rabbit hole.
It's brilliant.
We use Playwright MCP to automatically generate documentary for a website. I think it was Jeff's website probably yeah, yeah, and it did screenshots.
And now at hire MVP yep.
Everything.
It gets the logs, it gets the telemetry, it gets the errors. It can stop and restart individual resources and run commands against them, et cetera. And then play right's the one that's actually clicking for you? And it's like what.
But then on the other end, you know, I want to be able to add a resource into my app that's an AI resource, and so I want to be able to point it to a deployment on Foundry, which they have their whole new their like next gen Foundry experience that we were trying to build around.
Tell us about Foundry just really quickly. I mean, I think of it as a Microsoft alternative to a LAMA. But tell me what it is.
Officially, Foundry is. I hope no one on the Foundry heame ever listens to this. It's a hub.
For AI cloud apps, right, agentic apps, so you can get all your models, you can get all your very lightweight like data and stuff all there. It's a playground. The web experience came first. Foundry Local is their o LAMA thing. Yeah, yeah, and it's all this kind of one brand. Really, the point of it is to pull the complexity of Azure out of the experience because, as you all know, like if you're an Azure dev or someone who's an Azure admin, like there's a rabbit hole.
And so Foundry really streamlines it and makes it super easy for people who aren't already in the Azure ecosystem to hop in and start playing around. We also, like we have a GitHub models integration, We have just a generic open AI integration. You can stick any connection string anywhere into your spire app host. But having a really
streamlined Foundry experience is good for all you. And also, like Rich was saying earlier, it's good for us because it ties us to revenue, like we're an open source project, So we need to tie to revenue or when the contraction happens and the bubble starts to burst a little bit, we need to make sure that we are actually like increase and shareholder value.
Every CFO is going to go to every death and say show your ROI yep, And if you can show that you're providing value to the company, you live. And if you don't, you're gone.
Yeah, And it's a win win for Foundry because they don't have to worry as much about the devas, which is a really complicated problem right now, the devas of building AI apps. Like, yeah, together we've really been able to create a good story.
Well and just that awareness to put people in a place where they can demonstrate value.
That's good tooling.
You know, you can't have to remind people are supping Microft actually a tooling company. Yes, that's what it was really always about. Yeah, they happen to make languages and things along the way, you know, arguably against their will they had to, but they you know, and it's easy to focus on that.
Yeah, that's the business is tooling.
Yeah, I mean as fire was originally called a stack, which whole other thing, and it's it's the gottenet platform team that built a fire, like not the dot net tooling team, right, And at some point after I had joined, we kind of were.
Like, this is kind of a tool. Is this tool? And it totally changed our perspective on.
All of it.
Yeah, and then suddenly the languages don't matter and the platforms don't matter and it's just about being productive.
Yeah, yep.
And at the same and all the good stuff I see three different axis is moving at once. Right, that there is a shift of the platforms underneath you, a shift of the languages off of the side. And then suddenly this new wave of tooling with llms and things, and you need to hook into each of them. Like you look at this list of features, like why are they worried about this? It's like, because this is all the stuff that will stop you, yes, as a developer, from using this tool.
Yeah, exactly.
Yeah, that's really interesting. And what a battle.
Yeah, we have a constant you know battle, that's how much do we do and force ourselves to own versus how much do we let you fiddle around with? And so we still we haven't really come up with any strict guidelines yet, Like I think we're now we're talking about like, should we have install things for you if they're in your appost and you don't have them, Like does that make sense? Or should we point you at
something like is that too heavy handed? But for us, it's just you should be able to break out of the of our opinions at any time, like no black boxes is our number one thing.
That's exactly the thing I was thinking. You have opinions, but so do your customers. Yes, and so it's always a question of whose opinions are more tightly held.
Right right in our given moment, loosely held but we think they're good.
Yeah, but you can you can pop out at anything.
So I'm not sure what to do. The fact that you have an opinion makes my life better. But if I haven't told that's the way I'm doing things, you need to back off on your opinion or I have to go somewhere else.
Yep, exactly.
So the Greenfield story is pretty obvious. You start with a new visual studio template, you say I want an Aspire template, and then you go from there. Even if you're not using any of the caches and all that stuff, you could still just start that way and you get a little bit of extra stuff like the startup app, and then you have to click your app to run it.
But it's not it's not going to impede anything, right, so you should always start a new project with aspire I. Yes, but the brown Field thing is a little bit different, isn't it. And then this is what we did on stage that we were talking about. We aspiified dot net Rocks, the dot NetRocks dot com website, and it was pretty freaking easy.
It's it's it's really a manifest, like it's a type safe manifest. There are things you have to wiggle your way through sometimes, and especially with older dot net apps. So we had Handsleman come on and do it a spiification of his podcast website too, and and you know, he's using the same builder pattern that he used in dot net Core and we're like, bro, like, doesn't work
like that anymore. So there's definitely this modernization that has to happen because one of our opinions, right is we're not gonna let you keep using the old stuff, because we don't want the tech det of supporting all the old stuff and every matrix of that. We would never be a to get anything done if we had to worry about you know, the original ihose builders.
So you are forcing me down the pit of modernization here.
Yeah, but if I remember correctly, it was just like a right click on a solution or something, and it was like ad Aspire or was that easy?
It's that easy.
And now we have a command line experience too, so we'll look and we'll a knit and we're starting to build out for multiple languages. So if you have a job strip front end and dot Net back end, we can aspireify that from the command line, from VS code or from vs without you having to jump through hoops.
So I did a I have a customer that we hire Jeff Fritz to aspireify their application, right dot net Blazer application, and he did it and it worked fine. But there was only one thing that got in the way, and that was we were using poly the in it, and then you know, Aspire wants to use the poly stuff that's in the dot Net framework, but you know, just different name spaces and things like that, so that became a little bit of a tweak that we had
to get around. But other than that, it was just like easy.
Yeah, we turned It's one of another one of those things we turn on because we think you should probably have some level of resiliency.
Yeah, but you can just comment that out.
Like, that's the reason that the service defaults, which is our our opinions for a dot net back end or a dot net app, whatever it is. That's the reason it's a template and not a package you pull in so that you can go into it and comment things out.
It was very easy, Yeah, easy to upgrade. So good names all around Greenfield, Brownfield, yeah, blue Field. Yeah.
The other thing, like we keep talking about, nobody's going to make an Aspire app that's not like a thing.
It's an app that uses a spire.
So you'll do NPM create or you'll clone a template, or you'll go into visual Studio and you'll do file new project. And like, aspires should be easy to add in at any stage.
I can tell you the other side of this is coming into a project as a consultant and someone says, we used Aspire for this thought. I know a bunch of stuf about you know, hmm right, I know a whole bunch about this app. Things you will do and won't do if Aspire is actually working.
And you know, it'll probably be running on your machine that day instead of in two weeks.
Yeah, but you know, but they never call you when there's everything's working.
Well, that's true.
That's a good point.
That's a great point.
So so where are you struggling? Well, Maddie, it's been wicked awesome talking to you. And is there anything sorry? And is there anything else a that we missed that you want to I don't think so.
I mean it's going to be a crazy year. I think is going to be a wild ride hopefully.
I see you guys at MVP SUMM.
You know you're doing the two thousandth episode. That'll be great.
Yeah, I know.
I've got a I've got a special guest list to make sure folks can get in there because it's going to sell out.
So you're on that list, my dear, Oh I will.
But yeah, I think it's going to be a fun year.
We're hoping to do an Aspire centric online event, like similar to a dot net conf but about aspire'd be great at some point, So keep your ears open. I might tap you too for that. As the dot net heads, maybe you can come in and represent the old Guard.
It would be the token old dot net guys.
Yeah, really good for the for the stereotypes.
Yeah, I think you need a Python time when you need us.
Yeah, have someone in mind.
I'll circle back with you.
Of that, everybody has to be wearing depends. That's the yes, that's the me too.
On the outside, the streams can get long, you know, I get it. But yeah, Aspire dot dev is the home for everything Aspire, right, so that's the fight now, I love it, So just go there.
We stream on Fridays. We're doing more content. We're around. We're pretty honest to a fault. So we have a discord. It's all link on the website coming.
Out And Maddie, what's your home base on the web these days?
Probably Blue Sky is the thing I use the most, although I now I get to impersonate Aspire. I'm logged in as that brand account, which is horrifying.
Wow, but a god, you have a brand account.
Look at you as is here.
At Aspire dot dev the website or at Mattie Montacuola dot net is my personal one, which I use but mostly just to retweet other people or heckel Fowler.
So great and we all know you need to heckle Fowler more.
Yeah, yeah, all right, Maddie. We'll catch up with you later this year. Thanks. See you guys, all right, thanks and we'll talk to you next time. Rock dot Net Rocks is brought to you by Franklin's Net and produced by Pop Studios, a full service audio, video and post production facility located physically in New London, Connecticut, and of
course in the cloud online at pwop dot com. Visit our website at d O T N E t R O c k S dot com for RSS feeds, downloads, mobile apps, comments, and access to the full archives going back to show number one, recorded in September two thousand and two. And make sure you check out our sponsors. They keep us in business. Now go write some code, See you next time.
You got jamdvans
And