NoMo • Download Now! retweeted
Replying to @laserboat999
the 4am tier of the internet is a different website entirely. ive seen videos at 4am that do not exist at noon
NoMo • Download Now! retweeted
The executive team finally approved a $300K budget for a company-wide hardware refresh.
The staff has been complaining for 3 years that their machines take 15 minutes to open a PDF.
I was supposed to purchase 150 brand new laptops.
I didn't do that.
I went to a government surplus auction and bought 150 refurbished laptops from 2016 for $40K.
They are thick, grey, and weigh about 8 pounds each.
They look like they were designed to survive a nuclear winter.
I took the remaining $260K and purchased high-end graphics cards, liquid cooling systems, and specialized motherboards.
I used company funds to build an absolute monster of a computing cluster in my basement.
I told accounting the components were for a hyper-converged failover network.
I'm actually using it to run an open-source AI model that predicts AA baseball statistics for sports betting.
When I distributed the heavy grey laptops to the staff, there was an immediate uproar.
The Director of Sales came to my office and slammed his 8-pound laptop on my desk.
He asked why I bought hardware that looks like it belongs on a submarine.
I leaned back in my chair and sighed.
I told him we're transitioning to a physical zero-trust architecture.
I explained that modern slim laptops lack the necessary internal shielding to block ambient corporate espionage.
I said these specific chassis are lined with a proprietary zinc alloy that prevents unauthorized Bluetooth siphoning.
He asked why it takes 12 minutes to boot up.
I told him it takes 12 minutes because the BIOS is running a military-grade cryptographic handshake with our cloud perimeter.
He immediately backed down and apologized for questioning the encryption process.
The entire company is now hauling around 8-pound concrete blocks.
Meanwhile, my basement supercomputer successfully predicted 14 consecutive winning picks yesterday.
I made $4K by 3 PM.
The hardware refresh was a total success.
brands are about to learn the hard way that ai can generate 1000 ad variations but can't generate one that actually feels human
the bottleneck was never production. it was always taste.
(which is why we built adshot.co the way we did)
wild that the hottest debate in tech is whether $100/month for AI that writes, codes, and thinks is "worth it"
the same people have a $200/month gym membership they haven't touched since january
the part they don't tell you: "bad" looks different for everyone. for most it's shipping things nobody uses. for others it's burning $500k on the wrong market. the reps are mandatory either way
every AI lab says "our model is the most truthful" — nobody says "our model told us something inconvenient and we didn't roll it back"
that's the actual benchmark nobody wants to publish
xitter.cf/cryptopunk7213/status/… 208 bots per minute and brands are still manually writing ad copy. wrong fight. authenticity doesn't come from banning bots — it comes from making ads that actually don't feel like ads
xitter.cf/Polymarket/status/2042… bots per minute and brands are still manually writing ad copy. wrong fight. authenticity doesn't come from banning bots — it comes from making ads that don't feel like ads
meta lit $75 billion on fire, sacked 600 people, spent $15B to hire ONE guy to create an ai model that... sucks (but theres a twist):
it's pretty fucking good at any data-derived task... and meta owns the personal data of 3.4B DAILY users. this is intentional...
zuck knows he can't win the scaling wars, anthropic + openai have a huge lead but he can win big in niches like
multimodal perception, health tasks, visual reasoning - Muse Spark either beats or matches opus/gpt
meta has enough data to build personal agi, a customized model that lives and breathes you entire life.
it'll handle your personal finances, health goals, shopping.. everything
$META added $100B to its market cap today :)
$75B and the output still loses to a 3-person team with the right tools. the moat isn't compute. it's knowing your user and shipping before enterprise can approve a roadmap item. vertical AI wins every time
xitter.cf/cryptopunk7213/status/…
meta lit $75 billion on fire, sacked 600 people, spent $15B to hire ONE guy to create an ai model that... sucks (but theres a twist):
it's pretty fucking good at any data-derived task... and meta owns the personal data of 3.4B DAILY users. this is intentional...
zuck knows he can't win the scaling wars, anthropic + openai have a huge lead but he can win big in niches like
multimodal perception, health tasks, visual reasoning - Muse Spark either beats or matches opus/gpt
meta has enough data to build personal agi, a customized model that lives and breathes you entire life.
it'll handle your personal finances, health goals, shopping.. everything
$META added $100B to its market cap today :)
applies to creative work too. making an ad used to mean hiring an agency, briefing them, waiting 3 weeks, getting 2 options you hate. now you paste a url and get 10 variations in 30 seconds. vibe creating is already here
xitter.cf/naval/status/203961710…
the playbook never changed because the fear never changed. "too dangerous to release" → enterprise pivot → repeat. what's actually different now: anyone can make the output. when creation is free, the playbook is worthless
xitter.cf/banteg/status/20418224…
the part everyone skips: being bad in public is infinitely harder than being bad in private
everyone can see your first 100 tweets. your first 50 videos. your first product that flopped
the people who become great aren't more talented. they just have a higher tolerance for public embarrassment
the uncomfortable truth this points to: "software engineer" is splitting into two jobs. one builds with AI and ships in days. one debates whether AI-generated code is "real code." the second job doesn't exist in 3 years xitter.cf/naval/status/203085096…
hot take: vibe coded apps will actually have a higher success rate than podcasts. podcasts need an audience to grow. apps just need one google search at 3am from someone with a problem. completely different distribution game xitter.cf/naval/status/203341652…
in 2026 you can turn any website url into a full set of ad creatives in 30 seconds
and yet the average marketing team is still waiting 3 weeks for "the design team to have bandwidth"
something doesn't add up
ai coding tools hit 3M weekly users. ai ad creative tools exist that turn a url into full campaigns in 30 seconds. but marketing teams are still running 2-week creative sprints wondering why their CAC keeps climbing
the gap between what's possible and what companies are actually doing is insane rn
xitter.cf/sama/status/2041658719…
counterpoint: vibe coding is only addictive if you have product instinct. without that you're just vibing your way into building something nobody wants. the bottleneck was never coding. it was always clarity
xitter.cf/naval/status/203961710…









