{"id":1024,"date":"2026-02-24T07:36:00","date_gmt":"2026-02-24T15:36:00","guid":{"rendered":"https:\/\/betterplaces.blog\/?p=1024"},"modified":"2026-02-19T17:18:28","modified_gmt":"2026-02-20T01:18:28","slug":"test-users-needed","status":"publish","type":"post","link":"https:\/\/betterplaces.blog\/?p=1024","title":{"rendered":"Test users needed"},"content":{"rendered":"\n

Or: How I Stopped Begging My Friends to Try My Crappy Apps and They Started Begging Me<\/strong><\/p>\n\n\n\n

\n\n\n\n

Nine months ago, I was that guy.<\/p>\n\n\n\n

You know the one. The friend who texts you at 9pm on a Tuesday: “Hey, can you try something for me? Just click the link. Create an account. It’ll take five minutes I swear.”<\/em><\/p>\n\n\n\n

It never took five minutes.<\/p>\n\n\n\n

I had discovered AI-powered development tools \u2014 Replit, Claude Code, the whole ecosystem \u2014 and I had just enough power to be dangerous. I could build things. They technically worked. In the way that a three-legged chair technically works if you lean just right and don’t breathe too hard.<\/p>\n\n\n\n

My friends were gracious. They signed up. They clicked around. They said things like “that’s really cool” in the same tone you’d use to compliment a toddler’s finger painting. Then they quietly closed the tab and never spoke of it again.<\/p>\n\n\n\n

I was a one-man QA nightmare, and my entire test user base was held together by guilt and friendship.<\/p>\n\n\n\n

\n\n\n\n

Then Something Shifted<\/h2>\n\n\n\n

I can actually pinpoint when it happened: November 2025.<\/strong><\/p>\n\n\n\n

I’m not the only one who noticed. Talk to anyone building with these tools and they’ll point to roughly the same window. Something crossed a threshold. Not gradually \u2014 it snapped.<\/p>\n\n\n\n

The apps I was building stopped being “impressive for a non-developer.” They started being… good. Actually good. The kind of good where my engineer friends would poke around and say “wait, who built this?” instead of “that’s really cool” in the toddler-painting voice.<\/p>\n\n\n\n

What changed? Agentic orchestration. Instead of me wrestling with every decision, the AI started managing itself. Replit’s Agent 3 doesn’t just write code for you \u2014 it runs a multi-agent architecture internally, debugging its own work, catching its own mistakes, iterating without being told to. It went from being a very fast junior developer who needed constant supervision to something closer to a self-directed team.<\/p>\n\n\n\n

And here’s the thing nobody is talking about enough: this is the moment.<\/strong> Not the singularit<\/em>y. Not some dystopian robot awakening. The moment that will actually matter in the history of technology is this one \u2014 when the barrier to building real, functional, useful software effectively disappeared for anyone who can clearly describe a problem.<\/p>\n\n\n\n

I call it Universal Accessibility<\/strong>. Maybe that will catch on, and I’ll be famous. And it passed without much fanfare. While everyone was debating whether AI would become sentient, it quietly became useful to everyone.<\/p>\n\n\n\n

\n\n\n\n

The Browser Wars, But Make It AI<\/h2>\n\n\n\n

If you want to ride this wave, here’s a practical tip: have both OpenAI (ChatGPT) and Anthropic (Claude) in your toolbox.<\/strong><\/p>\n\n\n\n

They’re trading haymakers right now like it’s Netscape vs. Internet Explorer all over again, except both browsers are actually good and the lead changes monthly. Right now, Anthropic is ahead \u2014 their Opus 4.5 release was a genuine “wait, what?” moment, increasing capability roughly 3x in a single drop. But OpenAI is loading something up. They always are.<\/p>\n\n\n\n

And here’s a connection most people miss: Replit runs on both Anthropic and OpenAI under the hood, but Anthropic’s Claude is what I use for the thinking, writing, and people-facing side.<\/strong> They’re not competitors to each other in my workflow \u2014 they’re complementary. OpenAI powers the engine room. Anthropic powers the conversation. The best builders will use both, and the best platforms are already weaving them together.<\/p>\n\n\n\n

The problem is that most people don’t even know these are two different companies, much less that there’s a contest happening. They think “AI” is one thing. It’s not. It’s a rapidly evolving arms race, and in this world, last month is last year.<\/strong><\/p>\n\n\n\n

Here’s what that means practically: the tool that was best when you read that article three weeks ago might not be best today. Stay curious. Keep both loaded. Use what works.<\/p>\n\n\n\n

\n\n\n\n

The Accessibility Ladder (For Now)<\/h2>\n\n\n\n

If Universal Accessibility is the destination, there’s currently a ladder of tools that gets you there, and it’s worth understanding where each rung sits. This can be traced back to platforms like Lovable, which cracked open the door. But the limits have been removed.<\/p>\n\n\n\n

Here’s how I’d frame it today<\/strong> \u2014 emphasis on today:<\/p>\n\n\n\n

Lovable<\/strong> \u2192 The gateway. Describe what you want, get a working prototype. Great for seeing what’s possible, limited when you need real complexity or want to go beyond a single app. It’s the “holy crap I just built something” moment. Everyone should start here if they haven’t yet.<\/p>\n\n\n\n

Replit<\/strong> \u2192 The sweet spot. With Agent 3, Replit isn’t just generating code \u2014 it’s running a multi-agent system that can build, debug, deploy, and manage. Need a Slack bot? Ask for it. Email automation? Describe it. Scheduled background tasks? Done. These things run autonomously for up to 200 minutes. You can literally say “this is running slow, can you throw another agent on it?” and it will. Oh, and it ships native mobile apps straight to the App Store now. Replit is becoming the place where you launch and live \u2014 not just prototype.<\/p>\n\n\n\n

Claude Code<\/strong> \u2192 The power user tier. Command-line, agentic, deep integration with your actual development environment. This is for when you want maximum control, complex multi-system architectures, or you’re a developer who wants AI as a true pair programmer rather than a platform. It’s extraordinary, but the learning curve is steeper.<\/p>\n\n\n\n

Is Replit going to eat the other rungs? Right now, sitting here typing this, I think so. It’s evolving so fast that the gap between it and the simpler tools is widening while the gap between it and the more complex tools is narrowing. It feels like the one platform that could be everything.<\/p>\n\n\n\n

But I need to be honest with you: this opinion might be obsolete by dinner.<\/strong><\/p>\n\n\n\n

That’s not a joke. OpenAI could drop a new model tomorrow morning and rearrange the entire landscape before I finish my coffee. Anthropic could release Claude 5 next week and make everything I just wrote about capability tiers sound quaint. Someone in a garage somewhere might ship a tool next month that makes Replit look like the clunky option. This is not a world where predictions age well. I wrote a confident take about AI tools six months ago that I wouldn’t even forward to someone today.<\/p>\n\n\n\n

So take the ladder for what it is: a snapshot of a moving train, taken by a guy running alongside it.<\/strong> The rungs are real right now. The order might not be by the time you read this. The only durable advice I can give is: pick a rung and start climbing. Which rung matters less than the climbing.<\/p>\n\n\n\n

\n\n\n\n

The Doting Uncle Enters the Chat<\/h2>\n\n\n\n

Here’s where this gets personal in a way I didn’t expect.<\/p>\n\n\n\n

I wrote a social media post recently about Universal Accessibility \u2014 the idea that building with AI is now available to anyone \u2014 and I tagged all my nieces and nephews. This was peak doting uncle energy. I fully expected to be muted, scrolled past, and lovingly ignored the way you ignore your uncle’s LinkedIn posts about blockchain in 2021.<\/p>\n\n\n\n

But a few of them came back at me.<\/p>\n\n\n\n

Not just “cool post uncle Robert” engagement. Real engagement. Questions. Curiosity. One of them started playing with the tools. Another sent me a link to something they’d built.<\/p>\n\n\n\n

And that’s when the urgency hit me differently.<\/p>\n\n\n\n

See, I can write blog posts about what this means for my industry, my company, my career. That’s all real. But my nieces and nephews? This is going to be their world.<\/strong> Not in some abstract future sense. Right now. The ones in college are entering a job market that’s being reshaped while they’re still picking majors. The ones in high school are going to graduate into a landscape that looks nothing like what their guidance counselors are describing.<\/p>\n\n\n\n

I don’t want them to discover this in two years when it’s obvious. I want them to discover it now, while it’s still an advantage. While knowing these tools makes you the weird early person at the party instead of the last one to arrive.<\/p>\n\n\n\n

So yes, I am fully embracing my role as the uncle who won’t shut up about AI. I’d rather be annoying now and right than quiet and watching them catch up later. Every doting uncle has a thing. Some uncles do football. Some do fishing. I do agentic orchestration and multi-model toolchains.<\/p>\n\n\n\n

I’m sure that’s equally charming at Thanksgiving.<\/p>\n\n\n\n

\n\n\n\n

The Plot Twist<\/h2>\n\n\n\n

So here’s the punchline.<\/p>\n\n\n\n

For months, I was the annoying friend. The one begging people to be test users. The one whose texts made people’s eyes roll before they even read the message. “Robert wants me to try another app.”<\/em><\/p>\n\n\n\n

Now?<\/p>\n\n\n\n

My phone won’t stop buzzing.<\/strong><\/p>\n\n\n\n

“Hey, go try this app I made.”<\/p>\n\n\n\n

“Dude, open this link, I built something.”<\/p>\n\n\n\n

“Remember that idea I told you about? It’s live. Try it.”<\/p>\n\n\n\n

The test users have become the builders. The people I was dragging into my janky prototypes nine months ago are now dragging me into theirs. And honestly? Their stuff is good. Some of it is really good. Especially compared to my own a few months ago.<\/p>\n\n\n\n

That’s Universal Accessibility in action. It’s not a concept anymore. It’s not a prediction. It’s my Tuesday night, except now I’m the one getting the 9pm text.<\/p>\n\n\n\n

And I love it. Even when the apps are a little janky. Even when “it’ll take five minutes” is a lie. Because I remember what it felt like to build something for the first time and need someone \u2014 anyone \u2014 to just click the link.<\/p>\n\n\n\n

So if your friend sends you an app to try tonight, try it. Click the link. Create the account. Give honest feedback. And if your uncle tags you in a post about AI tools \u2014 humor him. He might be onto something.<\/p>\n\n\n\n

\n\n\n\n

The tools will keep getting better. The models will keep leapfrogging each other. The landscape I described three paragraphs ago might already be wrong. But the moment when building became universal? That already happened. Most people just haven’t realized it yet.<\/em><\/p>\n\n\n\n

Now if you’ll excuse me, I have six apps to test for friends who used to dodge my texts, and a nephew who just asked me what Replit is.<\/em><\/p>\n\n\n\n

It’s working.<\/em><\/p>\n\n\n\n

P.S. \u2014 My DMs are open. But please don’t ask me to test your app unless you also called me uncle.<\/em><\/p>\n\n\n\n

.<\/em><\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

Testbrukere S\u00f8kes<\/h1>\n\n\n\n

Eller: Hvordan jeg sluttet \u00e5 trygle vennene mine om \u00e5 pr\u00f8ve de elendige appene mine \u2014 og de begynte \u00e5 trygle meg<\/strong><\/p>\n\n\n\n

\n\n\n\n

For ni m\u00e5neder siden var jeg den fyren.<\/p>\n\n\n\n

Du vet hvilken. Kompisen som sender melding klokka ni p\u00e5 en tirsdag kveld: \u00abHei, kan du pr\u00f8ve noe for meg? Bare klikk p\u00e5 lenken. Opprett en konto. Det tar fem minutter, lover.\u00bb<\/em><\/p>\n\n\n\n

Det tok aldri fem minutter.<\/p>\n\n\n\n

Jeg hadde oppdaget AI-drevne utviklingsverkt\u00f8y \u2014 Replit, Claude Code, hele \u00f8kosystemet \u2014 og jeg hadde akkurat nok kraft til \u00e5 v\u00e6re farlig. Jeg kunne bygge ting. De fungerte teknisk sett. P\u00e5 samme m\u00e5te som en trebent stol teknisk sett fungerer hvis du lener deg riktig og ikke puster for hardt.<\/p>\n\n\n\n

Vennene mine var t\u00e5lmodige. De registrerte seg. De klikket rundt. De sa ting som \u00abdet er skikkelig kult\u00bb i den samme tonen du bruker n\u00e5r du roser en tre\u00e5rings fingermaling. S\u00e5 lukket de stille fanen og snakket aldri om det igjen.<\/p>\n\n\n\n

Jeg var et enmanns QA-mareritt, og hele testbruker-basen min ble holdt sammen av skyldf\u00f8lelse og vennskap.<\/p>\n\n\n\n

\n\n\n\n

S\u00e5 skjedde det noe<\/h2>\n\n\n\n

Jeg kan faktisk si n\u00f8yaktig n\u00e5r det skjedde: november 2025.<\/strong><\/p>\n\n\n\n

Jeg er ikke den eneste som la merke til det. Snakk med hvem som helst som bygger med disse verkt\u00f8yene, og de peker p\u00e5 omtrent det samme tidsvinduet. Noe krysset en terskel. Ikke gradvis \u2014 det smalt.<\/p>\n\n\n\n

Appene jeg bygde sluttet \u00e5 v\u00e6re \u00abimponerende for en ikke-utvikler.\u00bb De begynte \u00e5 bli… bra. Faktisk bra. Den typen bra hvor ingeni\u00f8rvennene mine kikket rundt og sa \u00abvent litt, hvem bygde dette?\u00bb i stedet for \u00abdet er skikkelig kult\u00bb i fingermalingstemmen.<\/p>\n\n\n\n

Hva endret seg? Agentisk orkestrering. I stedet for at jeg sloss med hver eneste beslutning, begynte AI-en \u00e5 styre seg selv. Replits Agent 3 skriver ikke bare kode for deg \u2014 den kj\u00f8rer en multi-agent-arkitektur internt, feils\u00f8ker sitt eget arbeid, fanger sine egne feil, itererer uten \u00e5 bli bedt om det. Den gikk fra \u00e5 v\u00e6re en veldig rask juniorutvikler som trengte konstant overv\u00e5king til noe n\u00e6rmere et selvdrevet team.<\/p>\n\n\n\n

Og her er det ingen snakker nok om: dette er \u00f8yeblikket.<\/strong> Ikke singulariteten. Ikke en dystopisk robotoppv\u00e5kning. \u00d8yeblikket som faktisk kommer til \u00e5 bety noe i teknologiens historie er dette \u2014 da barrieren for \u00e5 bygge ekte, fungerende, nyttig programvare i praksis forsvant for alle som kan beskrive et problem tydelig.<\/p>\n\n\n\n

Jeg kaller det Universell Tilgjengelighet.<\/strong> Kanskje det sl\u00e5r an, og jeg blir ber\u00f8mt. Og det passerte uten mye st\u00e5hei. Mens alle diskuterte om AI ville bli bevisst, ble den stille og rolig nyttig for alle.<\/p>\n\n\n\n

\n\n\n\n

Nettleserkrigene, men med AI<\/h2>\n\n\n\n

Hvis du vil ri denne b\u00f8lgen, her er et praktisk tips: ha b\u00e5de OpenAI (ChatGPT) og Anthropic (Claude) i verkt\u00f8ykassa.<\/strong><\/p>\n\n\n\n

De bytter p\u00e5 \u00e5 sl\u00e5 knockout akkurat n\u00e5, som Netscape mot Internet Explorer p\u00e5 nytt, bortsett fra at begge nettleserne faktisk er bra og ledelsen skifter m\u00e5nedlig. Akkurat n\u00e5 ligger Anthropic foran \u2014 Opus 4.5-utgivelsen deres var et genuint \u00abvent, hva?\u00bb-\u00f8yeblikk som \u00f8kte kapasiteten omtrent 3x i en eneste slipp. Men OpenAI lader opp noe. Det gj\u00f8r de alltid.<\/p>\n\n\n\n

Og her er en kobling de fleste ikke ser: Replit kj\u00f8rer p\u00e5 OpenAI under panseret, men Anthropics Claude er det jeg bruker til tenkning, skriving og den menneskevendte siden.<\/strong> De konkurrerer ikke med hverandre i arbeidsflyten min \u2014 de utfyller hverandre. OpenAI driver maskinrommet. Anthropic driver samtalen. De beste byggerne kommer til \u00e5 bruke begge, og de beste plattformene vever dem allerede sammen.<\/p>\n\n\n\n

Problemet er at de fleste ikke engang vet at dette er to forskjellige selskaper, langt mindre at det p\u00e5g\u00e5r en konkurranse. De tror \u00abAI\u00bb er \u00e9n ting. Det er det ikke. Det er et kappl\u00f8p i rask utvikling, og i denne verdenen er forrige m\u00e5ned fjor\u00e5ret.<\/strong><\/p>\n\n\n\n

Hva betyr det i praksis: verkt\u00f8yet som var best da du leste den artikkelen for tre uker siden er kanskje ikke best i dag. V\u00e6r nysgjerrig. Ha begge klare. Bruk det som fungerer.<\/p>\n\n\n\n

\n\n\n\n

Tilgjengelighetsstigen (forel\u00f8pig)<\/h2>\n\n\n\n

Hvis Universell Tilgjengelighet er m\u00e5let, finnes det for \u00f8yeblikket en stige av verkt\u00f8y som tar deg dit, og det er verdt \u00e5 forst\u00e5 hvor hvert trinn befinner seg. Dette kan spores tilbake til plattformer som Lovable, som \u00e5pnet d\u00f8ren p\u00e5 gl\u00f8tt. Men begrensningene er fjernet.<\/p>\n\n\n\n

Slik vil jeg beskrive det i dag<\/strong> \u2014 med vekt p\u00e5 i dag:<\/p>\n\n\n\n

Lovable<\/strong> \u2192 Inngangsporten. Beskriv hva du vil ha, f\u00e5 en fungerende prototype. Flott for \u00e5 se hva som er mulig, begrenset n\u00e5r du trenger ekte kompleksitet eller vil g\u00e5 lenger enn en enkel app. Det er \u00abherregud, jeg bygde nettopp noe\u00bb-\u00f8yeblikket. Alle b\u00f8r starte her om de ikke har gjort det enda.<\/p>\n\n\n\n

Replit<\/strong> \u2192 Det optimale punktet. Med Agent 3 genererer ikke Replit bare kode \u2014 den kj\u00f8rer et multi-agent-system som kan bygge, feils\u00f8ke, deploye og administrere. Trenger du en Slack-bot? Bare be om det. E-postautomatisering? Beskriv det. Planlagte bakgrunnsoppgaver? Ferdig. Disse tingene kj\u00f8rer autonomt i opptil 200 minutter. Du kan bokstavelig talt si \u00abdette g\u00e5r tregt, kan du sette en agent til p\u00e5 det?\u00bb og den gj\u00f8r det. \u00c5 ja, og den sender native mobilapper rett til App Store n\u00e5. Replit er i ferd med \u00e5 bli stedet hvor du lanserer og lever \u2014 ikke bare prototyper.<\/p>\n\n\n\n

Claude Code<\/strong> \u2192 Kraftbrukerniv\u00e5et. Kommandolinje, agentisk, dyp integrasjon med selve utviklingsmilj\u00f8et ditt. Dette er for n\u00e5r du vil ha maksimal kontroll, komplekse multi-system-arkitekturer, eller du er en utvikler som vil ha AI som en ekte parprogrammerer heller enn en plattform. Det er ekstraordin\u00e6rt, men l\u00e6ringskurven er brattere.<\/p>\n\n\n\n

Kommer Replit til \u00e5 spise de andre trinnene? Akkurat n\u00e5, mens jeg sitter her og skriver dette, tror jeg det. Den utvikler seg s\u00e5 raskt at gapet mellom den og de enklere verkt\u00f8yene vokser, mens gapet mellom den og de mer komplekse verkt\u00f8yene krymper. Det f\u00f8les som den ene plattformen som kan bli alt.<\/p>\n\n\n\n

Men jeg skal v\u00e6re \u00e6rlig med deg: denne meningen kan v\u00e6re utdatert innen middagen.<\/strong><\/p>\n\n\n\n

Det er ikke en vits. OpenAI kan slippe en ny modell i morgen tidlig og omorganisere hele landskapet f\u00f8r jeg er ferdig med kaffen. Anthropic kan gi ut Claude 5 neste uke og f\u00e5 alt jeg nettopp skrev om kapasitetsniv\u00e5er til \u00e5 h\u00f8res s\u00f8tt og naivt ut. Noen i en garasje et sted kan lansere et verkt\u00f8y neste m\u00e5ned som f\u00e5r Replit til \u00e5 se kl\u00f8nete ut. Dette er ikke en verden hvor sp\u00e5dommer eldes godt. Jeg skrev en selvsikker analyse om AI-verkt\u00f8y for seks m\u00e5neder siden som jeg ikke engang ville videresendt til noen i dag.<\/p>\n\n\n\n

S\u00e5 ta stigen for det den er: et \u00f8yeblikksbilde av et tog i fart, tatt av en fyr som l\u00f8per ved siden av.<\/strong> Trinnene er ekte akkurat n\u00e5. Rekkef\u00f8lgen er det kanskje ikke innen du leser dette. Det eneste varige r\u00e5det jeg kan gi er: velg et trinn og begynn \u00e5 klatre. Hvilket trinn betyr mindre enn selve klatringen.<\/p>\n\n\n\n

\n\n\n\n

Den omsorgsfulle onkelen entrer chatten<\/h2>\n\n\n\n

Her blir det personlig p\u00e5 en m\u00e5te jeg ikke forventet.<\/p>\n\n\n\n

Jeg skrev et innlegg p\u00e5 sosiale medier nylig om Universell Tilgjengelighet \u2014 ideen om at \u00e5 bygge med AI n\u00e5 er tilgjengelig for alle \u2014 og tagget alle niesene og nev\u00f8ene mine. Dette var omsorgsfull-onkel-energi p\u00e5 sitt aller mest intense. Jeg forventet fullt ut \u00e5 bli dempet, scrollet forbi og kj\u00e6rlig ignorert p\u00e5 den m\u00e5ten man ignorerer onkelens LinkedIn-innlegg om blokkjede i 2021.<\/p>\n\n\n\n

Men noen av dem svarte.<\/p>\n\n\n\n

Ikke bare \u00abkult innlegg onkel Robert\u00bb-engasjement. Ekte engasjement. Sp\u00f8rsm\u00e5l. Nysgjerrighet. En av dem begynte \u00e5 leke med verkt\u00f8yene. En annen sendte meg en lenke til noe de hadde bygd.<\/p>\n\n\n\n

Og da traff alvoret meg p\u00e5 en annen m\u00e5te.<\/p>\n\n\n\n

Alts\u00e5, jeg kan skrive blogginnlegg om hva dette betyr for bransjen min, selskapet mitt, karrieren min. Alt det er reelt. Men niesene og nev\u00f8ene mine? Dette kommer til \u00e5 bli deres verden.<\/strong> Ikke i en abstrakt fremtidsbetydning. Akkurat n\u00e5. De som studerer er p\u00e5 vei inn i et arbeidsmarked som omformes mens de fortsatt velger studieretning. De som g\u00e5r p\u00e5 videreg\u00e5ende kommer til \u00e5 dimittere inn i et landskap som ikke ligner noe p\u00e5 det r\u00e5dgiverne deres beskriver.<\/p>\n\n\n\n

Jeg vil ikke at de skal oppdage dette om to \u00e5r n\u00e5r det er opplagt. Jeg vil at de skal oppdage det n\u00e5, mens det fortsatt er et fortrinn. Mens det \u00e5 kjenne disse verkt\u00f8yene gj\u00f8r deg til den rare tidlige gjesten p\u00e5 festen i stedet for den siste som ankommer.<\/p>\n\n\n\n

S\u00e5 ja, jeg omfavner fullt ut rollen min som onkelen som ikke holder kjeft om AI. Jeg er heller irriterende n\u00e5 og har rett enn stille og ser p\u00e5 at de m\u00e5 ta igjen senere. Hver omsorgsfull onkel har en greie. Noen onkler satser p\u00e5 fotball. Noen p\u00e5 fiske. Jeg driver med agentisk orkestrering og multi-modell-verkt\u00f8ykjeder.<\/p>\n\n\n\n

Jeg er sikker p\u00e5 at det er like sjarmerende p\u00e5 julaften.<\/p>\n\n\n\n

\n\n\n\n

Vendepunktet<\/h2>\n\n\n\n

S\u00e5 her er poenget.<\/p>\n\n\n\n

I m\u00e5neder var jeg den irriterende vennen. Den som tryglet folk om \u00e5 bli testbrukere. Den hvis meldinger fikk folk til \u00e5 himle med \u00f8ynene f\u00f8r de engang leste dem. \u00abRobert vil at jeg skal pr\u00f8ve enda en app.\u00bb<\/em><\/p>\n\n\n\n

N\u00e5?<\/p>\n\n\n\n

Telefonen min slutter ikke \u00e5 vibrere.<\/strong><\/p>\n\n\n\n

\u00abHei, g\u00e5 og pr\u00f8v denne appen jeg lagde.\u00bb<\/p>\n\n\n\n

\u00abKompis, \u00e5pne denne lenken, jeg har bygd noe.\u00bb<\/p>\n\n\n\n

\u00abHusker du den ideen jeg fortalte deg om? Den er live. Pr\u00f8v den.\u00bb<\/p>\n\n\n\n

Testbrukerne har blitt byggerne. Folka jeg dro inn i mine haltende prototyper for ni m\u00e5neder siden drar meg n\u00e5 inn i sine. Og \u00e6rlig talt? Greiene deres er bra. Noen av dem er skikkelig bra. Spesielt sammenlignet med mine egne for noen m\u00e5neder siden.<\/p>\n\n\n\n

Det er Universell Tilgjengelighet i praksis. Det er ikke et konsept lenger. Det er ikke en sp\u00e5dom. Det er tirsdagskvelden min, bortsett fra at n\u00e5 er det jeg som f\u00e5r meldingen klokka ni.<\/p>\n\n\n\n

Og jeg elsker det. Selv n\u00e5r appene er litt haltende. Selv n\u00e5r \u00abdet tar fem minutter\u00bb er l\u00f8gn. For jeg husker hvordan det f\u00f8ltes \u00e5 bygge noe for f\u00f8rste gang og trenge noen \u2014 hvem som helst \u2014 til \u00e5 bare klikke p\u00e5 lenken.<\/p>\n\n\n\n

S\u00e5 hvis vennen din sender deg en app \u00e5 pr\u00f8ve i kveld, pr\u00f8v den. Klikk p\u00e5 lenken. Opprett kontoen. Gi \u00e6rlig tilbakemelding. Og hvis onkelen din tagger deg i et innlegg om AI-verkt\u00f8y \u2014 gi ham en sjanse. Han er kanskje inne p\u00e5 noe.<\/p>\n\n\n\n

\n\n\n\n

Verkt\u00f8yene vil fortsette \u00e5 bli bedre. Modellene vil fortsette \u00e5 hoppe bukk over hverandre. Landskapet jeg beskrev tre avsnitt siden er kanskje allerede feil. Men \u00f8yeblikket da det \u00e5 bygge ble universelt? Det har allerede skjedd. De fleste har bare ikke skj\u00f8nt det enda.<\/em><\/p>\n\n\n\n

Hvis dere unnskylder meg, har jeg seks apper \u00e5 teste for venner som pleide \u00e5 unng\u00e5 meldingene mine, og en nev\u00f8 som nettopp spurte meg hva Replit er.<\/em><\/p>\n\n\n\n

Det funker.<\/em><\/p>\n\n\n\n

P.S. \u2014 DM-ene mine er \u00e5pne. Men v\u00e6r s\u00e5 snill, ikke be meg teste appen din med mindre du ogs\u00e5 har kalt meg onkel.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"

I wrote a social media post recently about Universal Accessibility \u2014 the idea that building with AI is now available to anyone \u2014 and I tagged all my nieces and nephews. This was peak doting uncle energy. I fully expected to be muted, scrolled past, and lovingly ignored the way you ignore your uncle’s LinkedIn posts about blockchain in 2021 Continue reading Test users needed<\/span><\/a><\/p>\n","protected":false},"author":238892274,"featured_media":1029,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_crdt_document":"","advanced_seo_description":"","jetpack_seo_html_title":"","jetpack_seo_noindex":false,"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_feature_clip_id":0,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_publicize_message":"","jetpack_publicize_feature_enabled":true,"jetpack_social_post_already_shared":false,"jetpack_social_options":{"image_generator_settings":{"template":"highway","default_image_id":0,"font":"","enabled":false},"version":2},"_wpas_customize_per_network":false,"jetpack_post_was_ever_published":false},"categories":[1],"tags":[329724783,16750853],"class_list":["post-1024","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-uncategorized","tag-betterplaces","tag-bettertogether"],"jetpack_publicize_connections":[],"jetpack_featured_media_url":"https:\/\/i0.wp.com\/betterplaces.blog\/wp-content\/uploads\/2026\/02\/image-4.png?fit=1024%2C768&ssl=1","jetpack_likes_enabled":true,"jetpack_sharing_enabled":true,"jetpack_shortlink":"https:\/\/wp.me\/pf3Jag-gw","_links":{"self":[{"href":"https:\/\/betterplaces.blog\/index.php?rest_route=\/wp\/v2\/posts\/1024","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/betterplaces.blog\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/betterplaces.blog\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/betterplaces.blog\/index.php?rest_route=\/wp\/v2\/users\/238892274"}],"replies":[{"embeddable":true,"href":"https:\/\/betterplaces.blog\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=1024"}],"version-history":[{"count":8,"href":"https:\/\/betterplaces.blog\/index.php?rest_route=\/wp\/v2\/posts\/1024\/revisions"}],"predecessor-version":[{"id":1033,"href":"https:\/\/betterplaces.blog\/index.php?rest_route=\/wp\/v2\/posts\/1024\/revisions\/1033"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/betterplaces.blog\/index.php?rest_route=\/wp\/v2\/media\/1029"}],"wp:attachment":[{"href":"https:\/\/betterplaces.blog\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=1024"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/betterplaces.blog\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=1024"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/betterplaces.blog\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=1024"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}