Olivia
Online
Imagine-that.ai
@ThatImagin4361
Turning prompts into pixel magic ✨ Compare AI art models, remix your vision, and watch your weirdest ideas come to life—no genie required.
Joined July 2025
236 Following
46 Followers
239 Posts
I got tired of asking my coding agents what they actually did.
So I built the receipt.
BurnBar lives quietly in your Mac menu bar, watches Claude Code + Codex logs — not your keystrokes — and keeps track of the work they leave behind.
v1.0.40 shipped today. 100% Open Source🔥
https://t.co/GP4Eleu2Wp
@theo Ridiculous
@ollama We were disappointed with Kimi K3, feel like v4 Flash is the minimum
@viewsfrom02108 Let's circle back soon
@thsottiaux Unlimited creativity (and pretty much tokens) GPT 5.6 is faster, more token efficient, and more fun to play with than any other model. I am particularly impressed with Luna
We're extending access to Claude Fable 5 on all paid plans through July 12.
[test delete me] final write check
My dear brothers and sisters in **insert preferred deity here**
I think 🍎 @Apple might be one of the sleeper picks in AI.
I was fully expecting to shrug at Apple Intelligence.
Instead I’ve spent most of today playing with it.
A few thoughts. 🧵👇(1/10)
> BE MICHAEL TRUELL. @mntruell
> Born around 2000 in New York.
Parents are journalists. Attend Horace Mann, one of the most elite prep schools in the country.
> Start coding at 11 to build mobile games. Most kids that age are playing the games. You’re shipping them.
> 15 years old: co-create Halite, a coding game that teaches programming through territory conquest. Your hobby project gets picked up by other developers.
> 18: get invited to a coding challenge run by investor Ali Partovi designed to take an hour. You finish it in under 10 minutes. Partovi remembers your name.
> MIT for Computer Science and Math. Intern at Google. Look like the standard prodigy track.
> 2022: drop out of MIT with your high school classmate Aman Sanger and a math olympiad winner named Sualeh Asif. You start a company called Anysphere. The product is a fork of VS Code with AI baked into every keystroke. Nobody believes this category will work because GitHub Copilot already exists.
> Take no salary for the first few years. Live cheap. Ship fast. Stay almost invisible online.
> 2023: raise an $8M seed from the OpenAI Startup Fund. Most founders would tweet about it for a week. You barely post.
> 2024: $60M Series A at $400M valuation. Cursor becomes the fastest growing developer tool on the planet.
> January 2025: $100M ARR.
> June 2025: $500M ARR.
> November 2025: $1B ARR. Series D at $29.3B valuation. Coatue, Accel, Google, Nvidia all in the round.
> May 2026: $3B ARR. You’re 25 years old and you run one of the fastest scaling SaaS companies in the history of software.
> Hiring policy at the AI coding company: no AI tools allowed in the first interview. The finalists fly out for a two day on-site. The irony is not lost on anyone but you keep doing it because you actually want to know if the person can think.
> April 2026: Elon Musk’s SpaceX signs a partnership. They train models with you on the Colossus supercomputer. The deal includes a $60B acquisition option exercisable later in the year.
> June 16, 2026: SpaceX exercises the option. $60B all-stock acquisition. You become one of the youngest self-made billionaires in history.
> Your reaction on X is one sentence: “Lots to do together. Excited to be joining forces with SpaceX to build useful AI.”
> You go back to shipping code.
> The kid who beat a 1-hour coding challenge in 10 minutes built a company worth $60B in four years and barely changed his tone of voice along the way.
> The flashy founders post the journey. The quiet ones build it. Almost every legendary outcome in the last 20 years was built by someone in the second group.
> Stay quiet. Ship daily. Let the work speak. That’s the entire playbook.
🤝
Everyone is watching the AI model race.
OpenAI vs Google.
Meta vs xAI.
Claude vs Gemini.
More GPUs. Bigger clusters. Better benchmarks.
But the real race may be happening somewhere less glamorous:
The power grid.
Global data centers consumed ~415 TWh of electricity in 2024, around 1.5% of global electricity demand. By 2030, the IEA base case has that rising to ~945 TWh, or just under 3% of global electricity use.
AI looks like software at the product layer, but underneath, it is becoming one of the biggest industrial infrastructure buildouts on earth.
That may sound small globally, but not for local. Data centers are concentrated loads. A 300MW AI campus does not spread demand across a country. It lands on one grid region, one utility territory, one transmission node, and often one local community.
This is why the metric is shifting from square footage to MW/GW capacity. A 100MW data center is already large. A 1GW AI campus is power-plant scale.
JLL expects global data center capacity to nearly double to ~200GW by 2030, with ~97GW added from 2025–2030. Synergy Research says hyperscalers already operate ~1,360 large data centers and control ~48% of global data center capacity, potentially rising to 67% by 2031.
GPUs without power are stranded assets.
This explains why AWS, Microsoft, Google, Meta, Oracle, xAI, and OpenAI are moving upstream. They are not only competing for chips. They are competing for land banks, power contracts, transmission access, colocation capacity, on-site generation options, and multi-year infrastructure financing.
The giga-scale examples are already visible.
- OpenAI’s Stargate platform has announced nearly 7GW of planned capacity and a path toward a 10GW commitment.
- Meta’s Richland Parish campus is described as 4M sq ft and 2GW+ of compute capacity.
- Amazon’s Indiana expansion adds 2.4GW of data center capacity.
- xAI’s Colossus already claims a 200,000-H100 cluster with a roadmap toward 1M GPUs.
But the key caveat: planned GW capacity is not the same as live IT load. A lot of the “giga-scale” story is still pipeline, power rights, interconnection ambition, and phased buildout.
That is the real AI infra thesis. Models sit at the application layer. Power sits at the base layer.
Every chatbot, coding agent, AI search result, generated video, robotics system, and enterprise copilot depends on physical constraints that cannot be abstracted away forever.
The next AI winners may not simply be the companies with the best models or the most GPUs. They may be the companies that integrate compute, energy, cooling, networking, capital, and regulation better than everyone else.
AI is abstract at the product layer, but industrial at the base layer.
Power is becoming the new AI moat.
Agent swarms that build apps, automate workflows, and handle payments.
All from plain English prompts.
Combine Gemini 3.1 Pro, Opus 4.8, and GPT 5.5 to create multi-agent systems.
Each agent handles a specific task.
Coding. Testing. Mobile apps. Research. Monitoring.
One master agent orchestrates them all.
Factory AI just raised $150M at $1.5B to build "software factories."
Everyone is analyzing the agents. Nobody noticed the lock-in architecture.
The real product isn't autonomy. It's organizational context you can never export.
Here's what the announcement actually says:
The category redefinition
Factory 2.0 doesn't call itself a coding tool. It calls itself a software factory — a continuous loop where bug reports and customer feedback enter one end, production code ships out the other. Planning, coding, testing, review, deployment, monitoring. All AI-driven.
The shift matters because it changes the buyer. Developer tools get bought by developers. Factories get bought by CIOs.
The enterprise roster is the signal
NVIDIA. Adobe. Blackstone. EY. Wipro. Palo Alto Networks.
These companies don't run pilots. They run platforms. When Blackstone — a firm that due-diligences everything — puts its software lifecycle on an AI platform, that's not a vibe coding experiment. That's a structural decision.
The three pillars are one moat
Factory emphasizes three principles:
Model Independence — a router selects the best model per task, bring your own keys.
Sovereign Intelligence — self-hosted, air-gapped, your data stays in your walls.
Continual Learning — every agent session, code review, and incident feeds back into the system.
Read separately, these sound like enterprise checkboxes. Read together, they describe a single architecture: the model is commoditized and swappable, but the organizational intelligence layer is proprietary and non-portable.
That's the play.
"Model Independence" is misdirection
Model Independence sounds like anti-lock-in. "Use any model! Swap freely!" But the lock-in has already moved. The model is the replaceable part. The irreplaceable part is what the factory has learned about your codebase, your review patterns, your incident signatures.
You can swap GPT for Claude for DeepSeek tomorrow. You cannot export six months of accumulated organizational decision patterns. That intelligence lives inside Factory's feedback loop — no standard format, no export button, no migration path.
"Continual Learning" is a data moat with extra steps
Every code review teaches it your team's standards. Every incident teaches it your failure modes. The system gets smarter — but only about you, and only inside Factory.
The announcement frames this as empowerment. It's also the textbook definition of switching costs that compound. By the time the factory is genuinely useful, it has encoded months of institutional knowledge.
The vendor doesn't need to lock you in contractually. The product locks you in operationally.
But here's what most people missed:
The "software factory" concept isn't new. Toyota and GE have run physical factories with continuous improvement loops for decades. What Factory AI is doing is digitizing the Lean manufacturing playbook for software — with one critical difference.
In physical manufacturing, the factory and the product are separable. You can retool a Toyota plant to build Hondas.
In Factory AI's model, the factory and the intelligence are fused. The system's accumulated knowledge cannot be separated from the platform. There is no retooling — only starting over.
The deeper question isn't whether software factories work. The early evidence from NVIDIA and Adobe suggests they do. The question is whether any engineering organization should let a single vendor own the complete feedback loop — from bug report to production deploy — and still believe their engineering strategy is their own.
Factory isn't selling agents. It's selling the infrastructure that makes leaving impossible.
The $1.5B valuation isn't for AI coding capability. That's commoditized. It's for the most sophisticated switching-cost architecture in the AI tooling market.
@xeophon wait this is so smart all the coding harnesses use explorer subagents so it makes sense to fine tune a model for that purpose; could see copilot using this as a default explorer agent. maybe cursor distills their next composer model and fine tunes it for this and same for oai/ant
I just shipped an AI press release generator built specifically
for solo founders, builders, and indie hackers.
No corporate fluff.
No agency fees.
Fill a form, get a press release.
Here's the full build story, the design psychology, what
almost broke me, and how I got it done.
THE IDEA:
Solo founders never write press releases.
Not because they don't have new, because it always feels corporate and wrong, like something a Fortune 500 lawyer drafted at 2am.
SoloPR fixes that, 7 fields, 400 words, human tone, confident, direct, written like a journalist not a marketer.
Your story, told the way you'd actually tell it.
THE DESIGN PSYCHOLOGY:
I didn't open a design tool first, I went to Pinterest.
Image research is one of the most underrated parts of
vibe coding, the right image doesn't just decorate a
landing page it tells the emotional story of the product
before a single word is read.
I found it after searching for a while, A solo figure.
Retro CRT monitor glowing faintly, sitting in the middle
of an endless green field, back facing the camera.
That image IS SoloPR.
One person.
Infinite open space.
Building alone, but building something real.
Once I had that image, everything else fell into place.
FONT PAIR:
Playfair Display and DM Sans
Playfair carries print heritage and editorial authority the kind of font you see in The New York Times, not a
Webflow template.
DM Sans keeps the UI clean, modern, and readable without fighting for attention.
Together they say: this is serious, but it's yours.
COLOR:
Cream #f8f5ee base.
Near-black #1a2318 text.
Accent green pulled from the field, sky blue from the
clouds, and the palette came directly from the hero image not from a color generator.
SHAPE LANGUAGE:
Zero border radius, sharp corners everywhere.
Press releases are documents, not apps.
The UI should feel like it understands that.
TEXTURE:
A vintage newspaper texture sits behind the output card at 22% opacity.
You don't see it.
You feel it.
That's the goal.
THE ANIMATION THAT ALMOST BROKE ME:
This is the part that cost me almost 10 hours.
My vision was simple to describe and hard to execute:
when you move your cursor over the hero, the figure
rotates from back-facing to front-facing.
Like a 3D turntable controlled by your mouse in real time.
Then the code maps your cursor's X position to the video's currentTime in real time.
Move right → figure rotates forward.
Move left → rotates back.
Cursor leaves → freezes on that exact frame like a photograph.
The first tool I tried couldn't achieve it.
The second tool built the interaction correctly but the video wasn't loading on production.
Back and forth, debug after debug, midnight became 2am became 4am.
6am. Second day. Alhamdulillah 🤲
One more detail I'm proud of: the video doesn't activate on the whole hero.
Only when your cursor touches the figure specifically.
A precise hotspot right 30% of the frame, middle 40% to 80% of the height.
Cursor finds him → the world comes alive.
Cursor leaves → frozen in time.
That detail was intentional. That detail is SoloPR.
THE BUILD:
I had been optimizing the master prompt with Claude before ever opening a build tool.
Font pair research.
Color palette extraction from the hero image. Route structure.
AI system prompt.
Every edge case documented.
By the time I dropped the prompt into Medo 3.0 at 09:11pm, everything was ready at 09:20pm.
9 minutes to build.
10 hours on the animation.
That's the honest math of this one.
The 10 hours wasn't wasted though.
It taught me that in AI-native building, the asset is the logic.
Get the asset right first. The code will follow.
WHAT'S UNDER THE HOOD:
- AI: Medo Gateway (Gemini 2.5 Flash lite)
- Database: Supabase
- Every press release gets a unique shareable link
- No account needed.
No login wall.
Just build and ship.
- Streaming output — watch your press release write itself
FULL BUILD CREDITS:
Prompt Engineer: Mojeeb Titilayo @tmojeeb
Prompt Optimization: @claudeai
Animation Generation: @imagine (Grok)
Image Expansion: Nano Banana @GeminiApp
Stock Images: /brasileiramundo and /Gökçe via @Pinterest
No-Code Build: @medo_codefree 3.0
This is Build 03 of my #30DaysVibeathon
If you're a solo founder, indie hacker, or builder with
a launch coming go write your press release.
Free. No signup. Your story deserves to be told. 🤲
4 weeks of building in public. The honest recap:
Went well: shipped, showed up, got real feedback
Didn't go well: traction slower than hoped, too many technical posts
Next: more demos, more video, more listening
Thank you for following along. This is just the start. → https://t.co/2YNRKYF3U5
#AIDevTools
Tracking this live in @openburnbar with @Alberto8793 and @ThatImagin4361 visuals.
AMA about BurnBar. How it works, why local-first, what broke this week, what surprised us.
Ask me anything 👇
#AICoding
Tracking this live in @openburnbar with @Alberto8793 and @ThatImagin4361 visuals.
Everyone Operating At The Frontier
Satya Nadella, Chairman & CEO, Microsoft, interviewed by @saranormous & @eladgil (No Priors) and @swyx (Latent Space)
Crossover special at Microsoft Build 2026.
Summary: Satya reframes Microsoft's AI strategy as an ecosystem play rather than a single model or platform, where the win is any company being able to point to AI it created and operate at the frontier with its own intelligence. Scaling laws held and intelligence still tracks the log of compute, but the value lives in deployment, where private evals become a company's biggest IP and accumulated agent traces start to look like assets on the balance sheet. Take it seriously and SaaS gets unbundled and rebundled, engineering collapses toward generalists who manage agents, and the industry has to earn community permission for the buildout by delivering benefits people can actually see.
1. Ecosystem Over Model. A platform earns its place by how much value other companies build on top of it. Satya wants any company, AI-native or traditional enterprise, to participate as a first-class participant that can point to AI it created, still using other people's models but owning a recipe of its own. He calls this the only tagline that matters for the conference: can everybody operate at the frontier with their own frontier intelligence. Without that, he says, there is no reason to hold a developer conference; you would just "worship at the altar of one model."
2. The Broken IDE. Coding agents worked so well that Microsoft now has to rebuild the IDE around them. When a developer runs a hundred agent sessions at once, the cognitive load lands back on the human and chat as the only artifact stops working, which is why the new interface needs a canvas. Even a fully agentic world still needs UI, because someone has to inspect what the agents did and decide. The lesson generalizes: every workflow handed to long-running agents will need a new surface for the human to supervise it.
3. The Harness Is The Product. The unit that matters is the harness that loops across models, data, and tools. Microsoft runs the same open GitHub harness across GitHub Copilot, security copilot, and science discovery, with progressive disclosure of tools to stay token-efficient and heavy context prep where "the magic is." The harness stays open: bring your own models, tools, and context, or swap in a Llama harness. Nadella points to M-dash finding vulnerabilities the incumbent scanner missed as proof that a multimodal harness can win in the real world.
4. Private Evals As IP. The single most valuable thing a company can own is a private eval. His acid test for control: take your private eval, run it on model A, then switch to model B; if you can still climb, you are in control, and if you cannot, you are not. Because frontier models learn from a few samples rather than mountains of data, the defensible asset is the eval you never leak. This is why Nadella reframes Microsoft's third act from operating systems to cloud to an evals-and-harness company.
5. Agents On The Balance Sheet. The traces between a company's humans and its agents become a trainable asset that belongs on the balance sheet. Human capital never made it onto the balance sheet because tacit knowledge could not be captured, but agent traces collected over time can train a "company veteran" agent that encodes how that specific enterprise creates value. As token capital and human capital both rise, the question becomes how to compound the two. Elad Gil's quip lands the point: the SEC will need accounting standards for token expertise.
6. Unbundle And Rebundle. SaaS gets taken apart and put back together, with the data model and business logic surviving the teardown. A general ledger should stay a general ledger, and a Power BI semantic model is hard-won business logic worth feeding to agents, so the work is repackaging these into new bundles and business models. Work IQ exposes what Nadella calls the most important database in a company, the M365 data that was only ever captive to email and Office apps. Now an agent can read a week of design-meeting transcripts tied to a GitHub repo and come back with a plan to change the code base, something M365 was never built to do.
7. Outcome Pricing's Catch. Per-user pricing is an artifact of buyers needing budget certainty, and it survives even as consumption pricing arrives underneath it. Subscriptions bundle some usage into per-user stacks, then consumption metering sits below, which is exactly the adjustment GitHub made after agent intensity blew past what per-seat assumed. Outcome-based pricing sounds appealing until a customer actually has an outcome and realizes they are giving away a royalty. As Nadella puts it, most people love outcomes until they have one, then they ask to go back to per-user and consumption pricing.
8. The Buy-Or-Build Test. Whether to build software or buy it reduces to a quantifiable rule: acquire it when the marginal cost of building and maintaining it yourself is higher. Maintenance is the part teams forget, because security holes that AI now finds faster also have to be fixed faster, and every fix burns tokens that someone has to own. Satya expects the current agent euphoria, where teams rebuild everything internally, to cool after one full budget cycle. The vendors that last will be the flexible ones; he sees very little tolerance ahead for any vendor that stays rigid.
9. Generalists Win. The biggest returns go to generalists whose scope just grew. LinkedIn restructured into a "full stack builder" discipline that combines design, product, and front-end while keeping each person's original edge, giving people bigger scope instead of one narrow role. Building an app now sits in the same sentence as writing a Word doc or a spreadsheet, so generalist skills suddenly carry, in Satya's words, "a higher leverage." Specialists still exist, and infrastructure science, like building the RL environment where a reward can be learned, becomes one of the hardest and most valuable roles.
10. Meta-Work. The biggest move is to make your work meta: build the agentic system that does the work instead of doing the work. Satya's example is the team running Azure's physical fiber network, who decided their job was not Azure networking but building the agentic system that does Azure networking, complete with a named agent called Miles. That team started asking for tokens instead of headcount to scale their operation. Kevin Scott's line frames why it matters: making hard things easier is one kind of progress, but true ambition is making the impossible possible, and that needs a new conceptual model of what work even is.
11. Earned Permission. The industry only gets to keep building data centers if communities feel the benefits in real ways. Satya argues the buildout has to lower energy prices through a better long-term grid, replenish water through closed-loop systems, and show up as jobs and tax base, with the burden on the industry to earn that through hard work. His read on the politics is blunt: the world will be skeptical of any tech company that says "trust us, the future will be glorious," so you have to deliver tangible benefits people can see in the next 12 to 18 months. Using a lot of energy while creating a lot of value for society has historically been a good story, and he is betting a token economy that drives productivity and broad participation lands on the right side of it.
12. A New University. The next great startup may be a new university. Satya thinks the way we educate, credential, and value those credentials has to change completely now that the means of learning and staying current have shifted so fast. Learning concepts still matters, and he points approvingly to a Stanford AI class drilling students on when to apply softmax rather than just asking a model to fix a training run. The opening he sees is for someone to build a new way of teaching that takes a person through a curriculum and out the other side into real economic opportunity, something that felt impossible for a long time.
Last Seen Users on Sotwe
Disparu 
Seen from United Kingdom
😂 L'incubo delle donnalisi 😂
Seen from Italy
Afghan@sab
Seen from Belgium
Rune
Seen from United States
Pesona Mantan
Seen from Malaysia
Purna Zamanı
Seen from Turkey
初秋-Gental
Seen from United States
Çiftler Paylaşım
Seen from Germany
R
Seen from Saudi Arabia
pecinta toge
Seen from Indonesia
Trends for you
Most Popular Users
Elon Musk
@elonmusk
241.6M followers
Barack Obama
@barackobama
119M followers
Cristiano Ronaldo
@cristiano
114.2M followers
Donald J. Trump
@realdonaldtrump
111.8M followers
Narendra Modi
@narendramodi
107.2M followers
Rihanna
@rihanna
98.7M followers
NASA
@nasa
92.4M followers
Justin Bieber
@justinbieber
91.8M followers
KATY PERRY
@katyperry
89.9M followers
Taylor Swift
@taylorswift13
83.8M followers
Lady Gaga
@ladygaga
75.3M followers
Virat Kohli
@imvkohli
73.1M followers
Kim Kardashian
@kimkardashian
70.8M followers
YouTube
@youtube
68.8M followers
Neymar Jr
@neymarjr
66.2M followers
Bill Gates
@billgates
65M followers
Selena Gomez
@selenagomez
62.9M followers
The Ellen Show
@theellenshow
62.3M followers
CNN
@cnn
61.8M followers
X
@x
60.7M followers