professional overthinker, amateur millionaire
currently teaching robots to do my job
so i can afford to have a job
if you made it this far, we're friends now.
been digging into Computers RH and this one is actually worth watching.
it's a PC-building simulator on Robinhood Chain where you build rigs, upgrade hardware, overclock, maintain your setup and earn robinhood:0xd421141b9d6afa274572a747e9a3fdd24ba8c400. the game is already live, and the Genesis collection is a 3,333 NFT free mint scheduled for october 15, according to the official FAQ.
here's where it gets interesting. the NFTs boost your in-game earnings, and there are multiple ways to qualify without buying anything. 1,889 spots are allocated to collabs, 1,111 to players and 333 to creators. when i checked, 314 creator spots were still available, with applications through https://t.co/vcySdEwnXH.
the actual game has more depth than just clicking buttons. matching PC builds can give up to +48% bonus, island upgrades add up to +36%, and extreme overclocking can boost hashrate by 50% for 24h, with a risk of crashing your rig. maintenance matters too, since dust, thermal paste and wear can cut your performance. referrals reportedly pay 5% lifetime, and staking with up to a 2.5x island multiplier is planned.
my approach would be simple: apply for the creator WL, play the game and compete for the player allocation. no reason to throw capital at it before understanding the economy.
but don't confuse a free mint with guaranteed profit. the NFT might end up being worth more as an in-game boost than as something you can sell.
still, a live game + free mint + several entry routes is enough for me to keep this on the watchlist.
Computers RH is LIVE. 🏝️
Official $CRH Contract:
CA: 0xd421141b9d6afa274572a747e9a3fdd24ba8c400
🔗 https://t.co/r6lTvVBEyS
At launch, 60% of the total supply was purchased through the dev buy and sent directly to the reward pool.
The reward contract will automatically distribute 10% of the current pool every day to players.
All in-game prices are fixed in USD and every purchase is made with $CRH
60% of every token spent in-game goes back into the reward pool, adding to the following day’s rewards.
The island is open.
See you in-game. 🖥️💜
@coinerhit@RobinhoodCrypto@vladtenev it’s all gonna play out. there are assets that are massively undervalued with catalysts that could send them much higher. like you said, there could be a huge move. but… we’ll see, we’ll see.
dug into 10 NFT projects getting attention on my timeline. put together where each one stands, how to get in, and what’s actually worth doing as of oct 10.
1. Good Vibes Club - Cloud Factory
https://t.co/ygUjD28AXv
stage: new collection announced, 3,333 supply on robinhood chain.
entry: GVC holders, Stoke Score, Cosmics, 1/1s and Stonkbrokers.
take: strongest brand in this list imo, but GVC floor is already around $600+. watch the access requirements and run the numbers before buying in.
2. Sockbrokers - Moonbirds × Stonkbrokers
https://t.co/OyM6jNh6iQ
stage: official collab announced, 10,000 NFTs.
entry: 5,000 reserved for Moonbirds Nesters. details for the rest haven't dropped yet.
take: just monitor for now. no confirmed entry path for outsiders yet.
3. Hollywood Hills
https://t.co/yex772P4RR
stage: film crowdfunding through Republic, NFT debut announced for oct 15.
entry: film investment starts at $200; Villa NFTs are tied to backing tiers.
take: the NFTs don't give you revenue share. this is a separate film investment play, not a typical WL flip.
4. RYFTCORP
https://t.co/r8aPlJmpZT
stage: NeuralAI migration to RYFT, with RYFT IDENTITY coming up.
entry: free WL application, planned supply of 4,444.
take: apply and track the distribution details. migration risk and potential supply dilution are the main concerns.
5. Loopia World
https://t.co/2ahNjRVUl6
stage: Season 2 of an already live MMORPG. Loopians applications are closed.
entry: application review is underway.
take: if you didn't apply, wait for the next opportunity. low priority for now.
6. IdentityUnits
https://t.co/KzPYRmNvWi
stage: AI agents create and curate NFTs. $UNITS is already live.
entry: original WL is closed. Genesis allocation is for IMD Seat Holders.
take: interesting AI narrative, but this is no longer a free WL play for newcomers.
7. FungoLabs
https://t.co/fPq2lmjC1b
https://t.co/CH0EQNgO5b
stage: early encrypted NFT project built around FHE.
entry: WL application with tasks, planned supply of 3,232.
take: one of the best free-entry opportunities in this list. interesting tech, but no money in until the contract is actually deployed and verified.
8. Stonkrobbers
https://t.co/Oqlr4DItk6
stage: early art teasers, no mint details revealed.
entry: unknown.
take: keep it on the radar, but there's no real reason to commit money yet.
9. zkunnamed - THE UNNAMED
https://t.co/uPvL1M3ZZb
stage: very early teaser phase, almost no public info.
entry: unknown. no mint details or official mechanics yet.
take: follow and wait for the first real announcement. pure speculation on potential community attention right now.
10. PlayCambria
https://t.co/UStuuC3Dkt
stage: Founding Charter sold out. testnet and the future $RSGP token are next.
entry: primary mint is closed, only secondary available.
take: wouldn't chase the floor just for a potential token. timing and tokenomics are still major question marks.
what are you guys grinding right now?
Introducing Sockbrokers, an official Moonbirds x Stonkbrokers collab.
10,000 Unique RWA PFPs on Robinhood burnable for physical socks.
5,000 Sockbrokers will be distributed to nesters in the OCG Collector Profile. More details soon(ish).
thanks gpt for the new interactive ui feature. i was talking about this six months ago, and it always felt weird that claude had something like this while chatgpt didn't.
this genuinely changes how you can use ai. it's now much easier to break down complex topics, brainstorm, do research, and build step-by-step plans.
one more idea for the future: custom gpts with configurable uis, the ability to connect your own apis, and build specialized apps directly inside chatgpt. basically, let people create their own tools for specific workflows.
but let's see where we are in 6-12 months. maybe we'll get there 👀
thank you! i've been improving hermes to fit my own workflow, and honestly, the last 4 days weren't wasted at all. all the architecture decisions, experiments and problems i worked through gave me a much better understanding of the whole process. i feel like i'm making better decisions now because of it.
i tried to build everything at once
or how i wasted days fighting free ai models
i started building my own ai harness with a pretty simple idea: write some skills, build a nice interface, and explore the space a little deeper. somehow, i went so far down the rabbit hole that i started designing my own memory os, combining the best ideas i could find in other projects. multiple memory layers, temporary and persistent memory, token optimization, context management, and a bunch of other things that were supposed to make the system smarter and more efficient.
then i started designing modes. they were supposed to be the main feature of my harness. the first one was going to be trench, a mode for finding promising early-stage projects, researching them, and discovering new opportunities. basically, i wanted to build a tool that would actually be useful to me and help me make money.
and that's where the shitshow started.
i was trying to build all of this using free NVIDIA NIM models. kimi k3 kept getting stuck waiting for the provider, followed by 429 errors. glm 5.3 ran into similar limits. nemotron 3 ultra turned out to be pretty unreliable for this kind of work: the model could generate a ton of code, only for everything to break because of an extra bracket or some other basic mistake. and that's without even getting into hallucinations, context loss, and the time spent fixing work that had already been done.
at some point, i realized i wasn't even working on the project anymore. i was fighting APIs, testing models, designing infrastructure, and endlessly waiting for another free model to actually do its job.
over the course of a few days, i got so deep into development that i basically disappeared from everything else. i stopped keeping up with the spaces i was interested in, lost touch with my usual activities, and spent a ton of time building infrastructure that hadn't even started providing any value.
and then i realized what my biggest mistake was.
i was trying to build everything at once instead of focusing on one thing. i wanted a complete system with its own memory, modes, skills, and all kinds of other features, when all i really needed at the beginning was one working tool that solved a specific problem.
and the funniest part is that i originally wanted to build a tool for myself. but somewhere along the way, i got so obsessed with designing the tool itself that i forgot why i started building it in the first place.
so the takeaway is pretty simple: focus on one target.
stop spending days designing perfect infrastructure. stop trying to implement every cool idea at once. stop building for a future that hasn't even arrived yet. build one thing that delivers real value, use it, and only then move on to the next one.
because you can spend forever building the perfect tool to reach your goal and still never get any closer to the goal itself.
and i guess that's the biggest lesson i've learned over the past few days.
$200k into Abstract, imaginary points and now the hunt for a drop. well fuck, what could possibly go wrong
let me say it straight: if i were these people, i’d already stop waiting for this to save them.
i haven’t followed Abstract at all for about a year now and i’ve deliberately avoided any interaction with this ecosystem. because from the outside, this whole thing has long looked like another scam pyramid, where the community keeps believing in the founder’s comeback, the next rework, the next quantum and another carrot being dangled in front of them for a few more months.
the biggest red flag for me came back on september 5, 2025.
that’s when a major investigation came out with allegations against @LucaNetz: the story with the gold watches, NFTs and other incidents.
https://t.co/fdBqkJcg10
then things got interesting.
Phin left. then more people started leaving the team. projects started moving to other chains or publicly considering it. at some point even the lead dev left.
at the same time, Abstract itself started looking more and more dead, while people kept waiting for that big comeback they’d been constantly hinting at.
and this is where i have one main question: if everything really is going that well, why have we seen such a consistent collapse over the last year?
you can blame the market, bad timing, competition and bad decisions all you want.
but when there are this many mistakes, while the team is gradually disappearing from the public eye, it personally no longer looks random to me.
it’s especially funny watching this now, when people are still discussing potential drops and future rewards after spending $200k+ on the chain.
meanwhile even Luca himself now seems as distant from Abstract as possible. he used to call himself co-founder, now that connection is barely visible anywhere. the Pudgy Penguins branding was removed from the positioning too.
i’m not claiming i know what’s happening inside the company.
but after the september 2025 story, the subsequent departures, the degradation of the ecosystem and the current attempt to explain everything with “legal told them not to post on x”, i’ve come to one conclusion:
people have been kept alive on hope for way too long.
and the worst part isn’t even Abstract.
the worst part is watching people keep throwing money into it waiting for a miracle because they’re promised another “soon”.
you don’t need a comeback anymore.
maybe it’s time to admit you bought a narrative that didn’t work out and stop feeding it with more of your money.
Over a month ago, I had a call with @DailaX about Abstract’s future. I kept it private because I saw too many people using their calls for engagement, and I didn’t want to do that.
—
Now every KOL who barely participated in @AbstractChain or only stuck around for two months seems to have an opinion. So here’s what I took from my conversation.
—
The call was relaxed and genuine. There was no desperate pitch to keep me or the @RealityCheck_RR community on Abstract. Diala listened to my concerns as a community member and a creator.
—
I was told the core team had been advised by legal to stay off X. I was also told the Penguin branding was removed because Abstract and @pudgypenguins are separate brands. On larger deals, I was told term sheets had been signed. These were real discussions, ready to launch at a moments notice. Not just hope or an announcement of announcement.
—
I was told the team had since been given the go ahead to get active on X and other social channels. I noticed a few posts from @CashBowie shortly afterward, though he seems to have stepped back again.
—
I can’t promise what comes next. But nothing about that conversation suggested the core Abstract team was giving up. They were early to a trend we’re now seeing take shape. I wrote about it last month in the article below.
—
I’ve been here, I’ve raised my concerns directly, and I remain bullish on Abstract.
@karanjagtiani04 yeah, that’s the direction i’m going in, making sure the interface doesn’t limit what the user can do. for now though, i’m mainly building it for myself. once i’m confident in the result, i’ll probably share it.
been sitting since midnight designing my own ai harness
customizable interface, agent modes, a full memory system, the whole architecture around the model
at some point i realized i’m basically building my own operating system for ai because existing tools don’t fit how i want to work
still a lot to fix, but this is probably the most interesting thing i’ve built so far
it’s mostly about the interface and how the functions are exposed.
Pi is built more for engineers — nothing extra, you figure everything out from scratch.
Hermes has basically everything, but localhost, desktop and CLI all have different functionality, and the memory is pretty barebones.
the most convenient harness for me is Codex Desktop. you can see what the agent is thinking and doing, it has the tools you actually need, and there’s very little to configure. honestly, it’s hard to beat for development.
the problem is that it’s basically locked to ChatGPT, and even with a project you can’t get every function working.
so I figured I’d rather build my own harness around the workflow I actually want and keep adding things as I need them.
50 people are following what i’m doing right now.
and i wanna remember this moment. i started my web3 journey back in 2024, but i’m still figuring out what i actually want to build, what i want to talk about, and where i want to end up.
i don’t want this account to be just about plays, research or another piece of alpha. i want to share what i’m building, my ideas, mistakes, thoughts and the journey itself.
there aren’t many of us here yet, and honestly, that makes this part even more interesting.
thanks for being here and reading my thoughts.