Piroska Rákóczi

@antoniolupetti AI engines prioritize sources with a "verifiable professional history." LinkedIn is the second most-cited domain in AI search.

@PiroskaR mmm... I suspect AI search relies more on semantic structure and consistency than on external profile signals😅

"Generative AI in depth." This paper outlines recent progress and how generative AI is evolving along multiple directions, in particular GANs, VAEs, and diffusion models. The core problem is: learning a distribution well enough to generate plausible new data. GANs learn through competition, often producing highly realistic results but with unstable training. VAEs adopt a probabilistic approach, offering more stability but typically less sharp outputs. Diffusion models start from noise and reconstruct data step by step, are slower but currently very effective. Each of these models represents a different trade-off between output quality, stability, and computational cost. Progress seems to move precisely in the direction of finding a better balance among these aspects. link.springer.com/article/10.118…

@Arden_2210 21

Don’t rush..!! the pattern isn’t what you think Drop your answer and test your IQ 🔥

@HeyAbhishek The AI generates. OK. But who will validate? essayty.com

Genspark just launched Build. • ideas into websites • ideas into apps • concept to prototype • prototype to code • powered by Claude Opus 4.7 A lot of tools help you generate ideas.

@letsgoelsewhere Yes. Trustworthiness is key. But how do you prove your trustworthiness? Put information on your page, the about page, that includes data from your LinkedIn profile, such as skills endorsements or badges from Facebook groups.

The shift is now explicit: → search → pathways → options → defaults AI trust networks: don’t rank sources they stabilise outcomes Once a pathway works— it becomes invisible infrastructure.

@khalidseo3 There is another thing. Trustworthiness. But how do you prove this? Put information on your page, the about page, that includes data from your LinkedIn profile, such as skills endorsements or badges from Facebook groups.

People want to show up in AI search, but they are still using old Google tricks. Here is a big secret. ChatGPT does not look at Google. It gets its live facts from Bing! A recent test showed a brand new hotel beating a famous one in ChatGPT. Why? The new hotel showed up more on Bing. If your brand is not in the top Bing results, ChatGPT will miss you. Read the simple facts here: khalidseo.com/bing-shapes-ch…

@thejustinwelsh Before you solve the problem test if it is big or small.

The biggest advice I'd give to entrepreneurs in their 20s: Become crazy obsessed with being an optimistic problem-solver. You'll fix big problems before pessimists finish talking about why it can’t be done.

@zuess05 No.

It used to take dozens of engineers to build a massive tech company. Now, one person with an AI subscription has the output of an entire startup. We are guaranteed to see the world's first one-person billion-dollar company soon. But here is the billion-dollar question Can a software engineer actually beat a TikTok influencer with an AI subscription to it?

@OmriBuilds Good idea. There is another one: validate your idea before start building.

Build the audience before you need it. Launching to 0 followers is playing the game on hard mode.

@JerryDukes5280 @TRUMP_ARMY_ Before moving anywhere, validate the idea on X. Are you interested how?

@TRUMP_ARMY_ What an idea! Where do you start? California, Colorado, Oregon, Washington, Illinois, New York? They would all be a great place to start. I think after you do the Blue you can expand to the Rest of the states. The FIRST ORDER OF BUSINESS is to get the Save America Act passed!💯🇺🇸

Is it time to dig into where the American taxpayers money's really going?

@launchx And before applying validate your strartup idea on X. Are you interested in the right method?

Got a brilliant business idea but not sure where to start? 🤔 Swipe through to see if AwardX Europe is the right fit for you! Apply by July 1st and win cash prizes, scholarships, and gain professional mentorship ✨

@TheWysestone @Jacksonsrule I know you were told, but it works. Validate you business idea before you start a business.

Well you're welcome to start your own business working 80-100 hours a week or even go out and spend 5-8 hours hunting and gathering food if you'd like. At some point you're going to have to find a way to survive. Personally I find it easy to go work a job where I work for someone else and don't have to worry about all the complexities that comes with being a business owner. I've watched every male in my family have their own business and have decided I much prefer time with my family and doing the things I like in my free time. You can act like it's so horrible but honestly we are living in one of the simplest times humans have ever experienced thus far.

People work 8-10 hours a day, 5 days a week for someone else’s profit. Then they spend 4 hours at night staring at screens Basically, living for weekends +4 weeks PTO on repeat until age 65. While the government takes 25-30% their income And we’re told this is “normal.”

@UWcontent I think you might validate your startup idea before starting a startup.

Gears E-Day wishlist: Shorter startup Gears 5 has a lengthy logo run at the start, and after almost 7 years it feels like an eternity to fire up daily

@Tengushee You can avoid much free work if first you valadite your startuo idea.

"We can't afford to do X with our startup budget... perhaps someone from the community wants to do it for us for free?" ... is the most exploitative bullshit. Staggered by how often it's gets suggested as a serious solution round here. Y'all blind? So they make profit via your free labour? If running as a for profit business then attempting to offset costs against user good will is clown shoes. Huge red flag.

@Sethmr I am working on a Chrome-extension that helps validation. If you are looking for an investor in pre-seed phase of you startup it is a real proof of traction.

@PiroskaR Nope, that's why I wasted my time building this: peanutgallery.live/watch But you learn and create so fast with Claude, spending time validating the idea might take as long as just figuring out a way to bypass the problem that used to exist.

🥜🥜🥜🥜🥜🥜🥜🥜🥜🥜🥜🥜🥜🥜🥜🥜 Peanut Gallery is live. 4 AI personas watching YouTube with you in a Chrome side panel 🎯 Baba Booey — fact-checker (Claude Haiku + Brave Search) 🔥 Cynical Troll — Groq Llama 70B, 120ms TTFT 🎧 Fred Norris — sound FX + one-liners 🤣 Jackie Martling — comedy writer 🎙️ chrome.tabCapture → Deepgram Nova-3 → sub-300ms transcripts. Rule-based director picks ONE persona per trigger, others cascade in like a real writers' room. ⚡ ~$1.15/episode BYO keys, or free on the hosted tier. 🔓 MIT licensed. One click the 🥜 on any YouTube tab. 📺 Demo: youtu.be/UcpUBcp8TRc ➕ Add to Chrome: chromewebstore.google.com/detail/peanut-… 🔗 peanutgallery.live 💻 github.com/Sethmr/peanut.… 🥜🥜🥜🥜🥜🥜🥜🥜🥜🥜🥜🥜🥜🥜🥜🥜

@bargava I know how to start. Validate your idea.

Over the last couple of months, I've met/talked/chatted with a fair number of 35-45 y/o , mid-level roles in large enterprises - who all want to do startup, but aren't sure how to get started. Some of them have deep domain expertise, some with some extremely deep industry connects etc. YC - but for mid-level managers. Curious why YC wouldn't want to try such a cohort out for once?

@Sethmr Have you validated the idea before coding?

I loved making this project. I'm a fan of startup culture and the podcasting space in general. Claude has taught me how to use .md files to make almost any project possible. When bugs are the issue, we need more logging. When you are learning a new tool, take snapshots everywhere. When you feel complete audit and ask for improvements. Constantly maintain and improve your local documentation from the AIs context and you can build anything!

@benkellyone If I were open to your offer, first I would validate the idea by talking about the problem your small business solve. If it is a comman problem, I would DM you. You might be great, but business starts with validating the idea.

If you want to learn exactly how you can purchase your first small business with little to no money down... DM me "SMB" I'll take you through my step-by-step strategy to help you acquire a lucrative business.

Imagine starting a business when you could just: • Find a Boomer with a boring business • Buy it with little to $0 of your own • Keep the leftover cash flow This strategy pays me $900k/year. Here’s exactly how:

@Mike_Scully_ You have done this well. First validate your solution and build after you gain attention. No homepage. No tool. No ads. Marketing starts long before building anything.

If I had to start again tomorrow with no followers and no clients, this is exactly what I’d do. No fancy tools. No audience. No testimonials. Just raw execution. I’d open a blank page and write down one thing: → What problem can I solve better than most? Then I’d turn that into a simple offer: “I help [type of person] solve [specific problem] using [my skill].” That’s the foundation. Then I’d do this for 30 days straight: ✅ Post daily, not to go viral, but to attract the right people ✅ Start 20+ real conversations per day ✅ Document wins, lessons, and results as I go ✅ Offer my help to 20–25 people every day ✅ Stay lean, move fast, and ignore the noise No waiting for a website. No logo. No endless planning. Just a skill. A problem. And consistent action. People overcomplicate this game. They think they need a personal brand before they sell. Or 10,000 followers before they launch. You don’t. If you can solve a real problem, you’re in business. Even with zero audience. This is how I started. It’s how dozens of my clients started. And if I lost everything, it’s how I’d start again.

@matt__makes @DanKulkov Comment marketing. And marketing starts long before coding. Validate your idea first.

@DanKulkov what kind of marketing are you thinking of starting with?

fuck i guess i need to do that "marketing" thing