Nev

My first AI company died in 36 hours this is vlog #2

@iruletheworldmo Which one is your favourite strawberry man

if you’re not using these skills there is no hope for you. bookmark this. use them.

@TylerDurden Lmk if you ever do want to go deeper, you could do some cool things with your website and course data too I think. A thing I love to do is doing a full research dive into anything, sends me a report to read w/ links

Tell me why I need Claude or a clawd bot? I don’t think I need more than grok currently which is a search engine on steroids. Convince me otherwise and I’ll send you $500 in bitcoin.

@kylemachol hahaa I think the human in the loop is important personally, even when an electric guitar with 100x pedals is available, sometimes just a clean signal is better

@nevaaron I'm still a caveman with a human and a loop so Claude Code prompts some Ralph loops but mainly still looked over by myself

Creating a plan for a large build with AI, something simple that has proved to be incredibly effective: "Audit the plan for any potential errors, issues or holes that could break once complete" First time, it found 19, next 15, then 5... keep going until you find 0.

@kylemachol What do you use to orchestrate? Openclaw? I do a similar thing with research questions, but gave them also separate clashing identities to encourage disagreements (and challenge each others thesis), but research is very different to implementation ha.

@nevaaron I feel like most models are trained to comply, so I always 1: spin up 3 and they have to agree on findings, 2. Generally have them penalized for false positives

@ashen_one Congrats man!! Cool milestone

If you wanna vibecode an iOS app, mine just got accepted to the iOS store There were a lot of stupid problems I ran into that I DO NOT want you to waste time on So read this (or feed to Openclaw), learn from it, and go make a million dollars

Not sure at all, it’s great for dramatic flare though. Being very real about it, the nasdaq is launching their own tech that will go onchain, Robinhood too. I think it might be the wrong question I’m answering (I also made this back in Sept) The financial system itself is changing, the traditional nasdaq might not exist as it has for the last however long, and become a different sort of exchange all together, and something more verified is 100% needed after phantom shares with GameStop etc, along with loads of other things. I personally don’t love perps tech that adds increase risk for people easily though. The ability to open up private markets that never leave private with series H’s though is great on the flip side of that

@nevaaron How sure are you about this

Hyperliquid will flip the NASDAQ

I use it to run multi model research swarms, enables different perspectives and approaches on a question and produces some really cool results I also use it pretty much exclusively for admin, have installed a lot of automated workflows like doing my accounting reconciliation via API I’ve built a dynamic to do list and idea log too, it autofills that etc That said, I rebuilt the entire memory framework with a full new brain basically hacking a couple of the root md’s for injecting different things, without that I think it’d be really stupid with endless bugs and bloat when trying to recall anything There’s a lot more little things I do as well, having multiple agents from different models is really cool though and provides really interesting results with lots of things

read many similar takes recently but curious to know what is it people are building w openclaw. I've tried coding, it's just bad unless you're working on workflow automation. the experience of building a website / tool w OC is frustrating, claude code / codex CLI is 100% the way to go. however, for automation, daily work and reporting it's been a huge help. personally haven't been encountering much of a problem w OC aside from coding.

@SCHIZO_FREQ @PD_EXP Oh boy do I need to talk to you about the project I’m working on

"I see you have a podcast lukas, what's it about?" "We attack the audience on a personal level for viewing our content"

Well! Man, this is a can of worms for me to answer haha. It's a topic I care deeply about, because the tech is so important but most of the tech has been taken hold of by scams and people who just want a quick buck, not using it to do what frontier tech is meant to do - make some sort of impact on the world (hopefully for the better). The money normally comes after you do the thing (innovate). It has meant that when you operate, you take 'signals' that you think are real, like X - but they aren't. It's been mostly bots, or people farming to create volume to get free money. Followers on accounts are mostly not real, and many transaction volumes are literal bot farms. They don't care about your product, but you think they do, and in the end you have no actual users because you were making choices off the wrong info (some of that throughline in the original post too - being careful about what you think is true, verifying from multiple places and not assuming things are because other people have) What I'd do: 1) Sit down and look at the product, and truly, truly, figure out, what problem does this solve? For who? Why does it exist? If not a problem, what value can it offer someone? 2) Once you know what the product does, for who, and ideally why. Make your strategy not just from a comms perspective (where, how) but also from a product perspective, and relentlessly focus on this If it's b2b, I'd stop wasting time doing wide reach campaigns on X or reddit, and focus on direct marketing and making real relationships with people that matter. There is likely a very small number. Kevin Kelly's 1000 true fans is an all time theory to read on this. Depth > breadth. I'd even change the product for them, likely. Most of crypto falls into this category right now imo. 3) If I sat down on these questions and realise the thing we are trying to create/sell has no problem that it solves, or could solve through iteration or a longer length of time - I'd stop working on it, or pivot aggressively into something that does. I'd then keep doing that until that happens, if the runway exists. This is the edge of technology, literal experiments, it is meant to fail by definition, to keep doing it and finding the thing that works. The places I am really excited about right now from the are not financial applications - those argubly have left experiment zone, it's payments, gambling and trading - bigger picture though, cryptography is the ying to AI's yang. Ironically, I think it will come out as the good guys, cryptography is one of the only things that can create rails not just financially, but in safety, security, privacy, to the unbelievable and literally unstoppable power that AI has the potential to become. So. I think, it's about finding ways the tech actually solves peoples problems in their lives, i think that will come from security, higher needs on maslow's hierarchy. Art has a place, too, verifying what is real and what's not, but the value isn't the fact it is has a cryptographic sig, it will be a requirement, the value will be the art itself. The people who certified things still made money though, just not in 'trading the certificates'. Idk if this helps but hope there's something maybe half valuable, I've been researching heavily lately the origins of the tech and how I think it has gone maybe a little to narrow focused on it's potential, hence having this length of an answer 😂

@nevaaron Wow Thanks for laying this out so well But what do you think about sectors like web3? It is assumed that most folks interested in anything web3 related are either on X or Reddit. Do you think web3 marketers should pay more attention to other social platforms?

Bit of marketing 101 Your surface area is your ceiling + you won’t ever hit 100% of a surface. If you’re trying to reach people, X is a small surface, full of people who might not need what you’ve made. Monthly active users (2026): - X ~570m - Instagram ~2.58b - TikTok ~2b - Facebook ~3.07b - YouTube ~3b X is literally 1/5th of Instagram and only 5% of these platforms combined! Those here are often also in the industry or ‘early adopters’, so will react with ‘I can just build it myself’. This represents a small group of potential people that also gives you a terrible barometer to make decisions on. Most people could learn or do lots of things themselves, but they don’t, it’s how it’s always been, there’s only so much time, and frankly, interest, otherwise we’d have never been a… society? The reality: most of your potential market are likely not here, and the ones you are desperate to impress (your peers) likely are… a very dangerous combo. Zoom out, it’s big bad ocean full of opportunity out there :)

@KiroIkigai These platforms are literally designed in every single way to make people feel that way

@nevaaron X is the smallest pond and somehow everyone is still trying to win it. 😄

@levithefirst I wish I took it myself sometimes, slapping myself in the face with this post tbh

@nevaaron zoom out good advice

@StarPlatinum_ I haven't!! Let's trade books Maybe not via mail tho, I'm in australia again now 😂

@nevaaron Have you read the blue ocean strategy bro?

Reading this thread, i just, man Saving money is as good as making money as far as I'm concerned. One of the biggest opportunities of AI is reducing costs Funnily enough, via decentralised means - building your own versions or open source. Contributing, etc. I don't know what this means for the world though, the current system at least. These products had 95% margins for years, the thought this is realistic, or even reasonable, is the problem to start with

@big_duca they have no leverage tho. this isnt tax software, it's hosting markdown files 😭

bye bye mintlify

@SpiralEye_ Haha less is more

@nevaaron ouuuf! couple that with "Breaking" and you'll have a timeline full of tumbleweeds! Bravo Nev!

The ai slop murder of the siren emoji was swift

@shachikyoto Haha you to man, not everything ai generated is slop, but the people making the stuff that is might be becoming slop itself if they can’t discern the difference anymore

@nevaaron GM Nev. Hope you have a productive low slop filled week 🤝

@emilios_eth Haha yeah, we can do with a little less sensationalism in times like these imo, especially when you can hit a duplicate mutate x1000 mass produce everywhere button

@nevaaron Ironically it because the definitive alert for sensationalism

@GadgetLeo Nah, the use is for things that aren't an emergency, they are just gaming something to increase 'dwell time' which tricks the algo into it being valuable content It will get killed site wide soon enough but i don't want to wait

@nevaaron ah more war side?