Barry Winata

Building @workstak_co

Also, writing @ Alpha Bytes: an anti-hype guide on how AI is impacting all of us -

[Barry Winata @barrywinata]

I've been building something quietly for the past few months and today felt like the right time to talk about it.

It's called Workstak, and it's designed to be curated marketplace for workflow tools. But the marketplace part isn't really the point.

The point is: every tool ships with an:
> Execution Kit.
>The actual workflow.
> The system prompts.
> The step-by-step blueprint to go from "just bought this" to productive in a single sitting.

I built it because I kept buying tools I never used. Not because they were bad, but because nobody showed me how to make them work in my specific setup.

I'd buy something on impulse, open it once, and it'd join the graveyard of browser bookmarks I'll never click again.

Figured I couldn't be the only one.

Waitlist is now open.

[Barry Winata @barrywinata]

SEO is not dead. Far from it.

Most people think that GEO is the future (which may be true), but SEO is still as important today as it was 5-10 years ago. In fact, the more tailwinds GEO gets, the more crucial SEO will become.

I nerded out over the w/e on SEO + GEO as I was doing a bunch of stuff with @workstak_co as well helping out a few portfolio companies improve their domain authority :)

Here's what I found that underpins all of this:

> GEO is downstream of traditional search - the fact is that LLMs don’t generate answers from static training data alone, but they use RAG to query the web in real-time. So if your technical SEO is broken (like blocking crawlers, poor sitemap structure, rendering errors, and even response times), AI scrapers can’t discover or index your pages in the first place.

> Also, the way LLMs construct an answer is to fetch the top 10-20 search results from standard indexes (like Google) and summarize them. So traditional SEO ranking determines which sources make it into the “retrieval pool”. If you don’t rank in traditional search, you won’t be summarized or cited in generative answers.

> Having structured web pages still matters (a lot!). LLMs are actually lazy. Things like JSON-LD (you should read more on this) are still super important cause it tells search engines exactly what your data means. They prefer to cite websites that are structured logically, load quickly, and use clear data formats (like bullet points and tables). LLM crawlers rely heavily on these structured headers to extract details, and they’re more statistically likely to cite sources that offer structured and verified schemas

> E-E-A-T and topical authority still matter because the same signals Google uses to evaluate trust (Expertise, Experience, Authoritativeness, and Trustworthiness) are analysed by generative engines to weight the authority sources - who's is linking to you? Are you cited by other reputable sites? The authority-building work you do for Google SEO is the exact same reputation score the AI uses to decide if you are a trustworthy recommendation.

So it’s not really SEO vs. GEO - it’s actually two sides of the same coin...

[Barry Winata @barrywinata]

@garrytan fair. but what is the "new game"?

[Barry Winata @barrywinata]

I don’t think we’ll be building traditional "apps" (and companies) much longer.

We’ll be building tools.

We're going to be moving away from monolithic destination apps to modular, utility-first tools that diffuse directly into our existing workflows.

It's prob fair to say that AI has dropped the cost of software production to zero. That means we’re about to get flooded with micro-utilities.

But here's the catch: when building becomes free, the bar for quality and taste goes vertical.

This is why I think the future is going to be modular.

Neat stuff.

[Jack Zhang @awxjack]

We started @Airwallex in Australia and are part of the great story of this country building significant companies that have had a global impact in consumer web, SaaS, and fintech.

I don't see any reason it can't do the same in AI. 🇦🇺

I'd like to see that happen. If you're building in AI and based in Australia, apply now 👉🏼 https://t.co/qbaB7zLEeN

[Barry Winata @barrywinata]

This is why rote learning is dead.

High agency—the default state of taking action, building things, and finding a way through when the standard path is blocked—is the new 21st-century meta-skill.

You can't lecture your way into that.

[Barry Winata @barrywinata]

I honestly feel like the universities are selling a false sense of security.

Millions of students enter college thinking their degree is a shield. That by showing up and following the syllabus, their future is safe. It isn’t.

The implicit social contract of higher education has been unilaterally rewritten. And this isn't just a tough job market—it’s an indictment of the educational system itself: 🧵

[Barry Winata @barrywinata]

https://t.co/OujRKhzT3x

[Barry Winata @barrywinata]

The rest will become expensive finishing schools for rule-followers.

If you graduate with a 4.0 GPA but don't know how to prompt Cursor or build a basic workflow with Claude, you aren't prepared.

You're just slow.

[Barry Winata @barrywinata]

https://t.co/OujRKhzT3x

[Barry Winata @barrywinata]

Google's Omni is insanely realistic.

But this kind of creates a bizarre economic loop

The running joke is that you can generate Omni videos, dump them on YouTube, and let the ad revenue pay for your Omni API bill and even for the model itself.

Think about it:

> You pay Google to generate the video (Omni API).
> You upload it to Google (YouTube).
> Google pays you ad revenue (AdSense).
> You use that revenue to buy more API credits from Google.
> YouTube gets ad revenue which could offset training/inference costs

A perfect, closed-loop Google arbitrage haha.

[Justine Moore @venturetwins]

Omni from @GoogleDeepMind just dropped 👀

It's a big step forward in video generation when it comes to character consistency, world knowledge, and editing.

I've been testing it for the last few days - and I'm excited to walk through some of the key features + my clips 👇

[Barry Winata @barrywinata]

@venturetwins @GoogleDeepMind Time to start a YouTube channel. The flywheel effect is real. This model will pay for itself.

[Jason Shuman @JasonrShuman]

Why second time founders obsess about distribution

[Barry Winata @barrywinata]

@JasonrShuman Exactly!

[Barry Winata @barrywinata]

@leopardracer Noice. Great write up. Having personal brand is mandatory now.

[Barry Winata @barrywinata]

There's going to be an absolute explosion of new apps and tools over the next few years.

Which also means an exponential increase in founders fighting for the exact same attention.

I think this is a good thing. Here's why:

> When software creation costs go to zero, the bottleneck shifts from "can we build this?" to "does anyone actually even care?". Most tools fail not because the tech is bad, but because they're looking for a problem that doesn't exist.

> The bar being raised is a massive win for users. They get better tools, faster onboarding, and actual support. The whole shelfware issue might finally start to fixt itself if builders are forced to care about retention over hype.

> It’s also an unprecedented time to be in the arena. If you're building right now, don't worry about the sheer volume of tools launching.

Just focus on being the single best option for a very specific group of people.

That’s how you cut through.

[Solo Founders @solofounders]

Yasser is seriously raising the bar for what it means to be a bootstrapped solo founder.

[Colossal Biosciences® @colossal]

We built an artificial egg and hatched chicks.

The Colossal artificial egg has a bioengineered shell designed to breathe like a real egg. And because a South Island giant moa egg was ~80x larger than a chicken’s, this will be critical tech for bird de-extinction.

[Barry Winata @barrywinata]

Is anyone out there actually using Gemini 3.1 (High) for agentic coding?

[Barry Winata @barrywinata]

Right I get it. Students feel threatened. They feel hopeless. They think they're graduating into a market where AI takes all the entry-level jobs. But if you actually think about it, the reaction doesn't make sense. And it's not the students' fault. The problem is the education system itself.

[Barry Winata @barrywinata]

This video of Eric Schmidt getting booed at a commencement speech for talking about AI seems pretty straightforward on the surface.

Students feel threatened. They feel hopeless. They think they're graduating into a market where AI takes all the entry-level jobs.

But if you actually think about it, the reaction doesn't make sense. And it's not the students' fault.

The problem is the education system itself.

Right now, AI is becoming the new politics. You bring it up, and people immediately retreat to their tribal corners. They're either wildly in favor of it or wildly combative. And plenty of universities/colleges are still operating with a deeply anti-AI mindset.

Without realizing it, they're actually setting a verry dangerous precedent. They're creating an "us-versus-them" mentality where AI is the enemy you have to hide from, rather than a lever you can pull.

Instead of coddling students and trying to "protect" them from AI, universities should be doing the exact opposite. They should be giving them high agency to exploit the hell out of it.

Students are literally the best demographic to take advantage of AI right now. They have the time, the blank slate, and the incentive to figure it out faster than anyone else.

Teaching them to fear it isn't protection. It's just really bad preparation.

[Alex Kantrowitz @Kantrowitz]

This is incredible. Artificial intelligence getting booed out of the stadium in any commencement speech it’s mentioned. Maybe telling college students AI was taking their jobs wasn’t the best strategy. Must watch —>