The firm mopping the data center floors is making more money than half the AI startups training models there.
November 2025 · 94 bookmarks
The firm mopping the data center floors is making more money than half the AI startups training models there.
Hottest Opportunity: Data-Center Cleaning
1. Largest revenue pool
2. Fastest growth
3. Still dirt-cheap
The dirtiest job in AI might be the best business.
You guys if this is slop then I like slop
All major neural networks, in one chart.
v/The Asimov Institute
Latest count: China now has over 150 humanoid robotics companies. We’ll definitely add more to our humanoid maps later.
it’s incredible how much better people get at math once you start explicitly mapping the colors out for them
Transformer by hand✍️
@AlexandraBalwit Reads like a David Foster Wallace essay lol. We'd be so lucky
Per Graphite, 52% of the articles written online are now written by AI.
The Slopocalypse is here.
excellent
Stuart Faught has been acquired 18 times on @acquiredotcom.
No plans to build a unicorn. Just trying to escape a corporate job that felt like Office Space. One tiny product at a time.
Dentists.
Orthodontists.
HVAC.
Med spas.
Home care.
Build vertical SaaS. Validate fast.
Sell around 50k ARR when it stops being fun.
Repeat.
His rule: “If a buyer’s a jerk, I’m out. If they’re cool, I stay flexible.”
Documentation. Niche clarity. Respect.
That’s how you get 18 exits on acquire.com.
Now he’s building Covington.ai with the same playbook.
Full episode:
"No one ever made a decision because of a number. They need a story." – Daniel Kahneman
@FeltInTheHeart Spoken like someone who has absolutely no idea how much power the media has over your hiring, sales, and suppliers. If you actually care about accomplishing anything in the long run, you have to push back.
The secret to public speaking is to speak as if you were alone.
🚨 SAUCE ALERT🚨
assuming you got a killer mobile app idea
how would I go to ship an MVP in less than a week as a non-coder/non-tech guy
1/ make a prototype with a vibecoding app
a few one that are famous
@rork @v_computer @anything
here the challenge for non-tech people is their struggle to explain with words what they have in mind.
which is basically what ai programmers do naturally because that's how our brain got used to think:
those vibe coding apps are great at coding the front-end screens
try to make at least a quick onboarding + 1 mvp feature
for the onboarding length, depends on your app niche + wonder this
"what would I need to see during the onboarding as my app prospect to be 1000% convinced I need that app to solve my pain point?"
the reason those health/fitness apps do 50/60/70+ onboarding screens is because we have to write a few sentences max per onboarding screen to not surcharge your users' brain
l- ong paragraphs kill conversions; micro-copy boosts completion rate, personalization accuracy, and trial starts
for the screens to look good make sure to brainrot yourself testing a lot of successful app + looking at apps' screens from @screensdesign_ to get a sense of what is a good screen design + a good onboarding flow for your app niche.
make sure to use one with which you get access to generated source code.
that's important for the next steps.
estimated time: 1 half day
this is only to get front-end screens designs
2/ if you had a terrible time with the previous step and you hate anything technical... work with someone for that part.
from here either get your hands dirty or hire a dev freelance or partner with a tech guy.
open a new project window with cursor or windsurf or similar using the source code you got
using your prototype as basis, you'll have to go through each app screen + asking llm to code the backend for your app (database, etc)
add a remote-controlled paywall at the end of your onboarding with superwall or revenuecat
estimated time: 3-4 days
Love seeing how iPhone features like 8x Zoom, Center Stage, and Dual Capture helped woodland friends create a musical masterpiece!
Shot on iPhone 17 Pro and directed by Mark Molloy, it’s a great reminder that kindness brings us all together.
Holiday cheer, a little bit of magic, and a whole lot of heart, captured on iPhone 17 Pro by a few handmade puppet friends!
Jeff Bezos on his favorite interview question
“When I interview people, I ask them to give me an example of something they've invented. And I always point out, it doesn't have to be something that you actually took to the patent office. It could be a metric that you invented and followed carefully. It could be a business process that you invented. You want to select people who like to invent their way out of boxes.”
A lot of people will immediately jump to what Jeff calls an “either or” solution (e.g. “We can do A or B.”). But as Jeff explains:
“The right question is how can we do A and B? What invention do we need to be able to do both?”
“Always stick to what makes you weird, odd, strange, different. That’s your source of power.”
— Robert Greene
Richard Feynman literally explained why smart people fail first:
⚡️The craziest prediction that is also the most likely:
Within 15 years, individual human cognition will stop being the primary driver of civilization.
Not “AI will get smarter.”
Not “automation will replace jobs.”
Not “education will be disrupted.”
I mean a total inversion of the cognitive hierarchy.
Here is the actual trajectory:
1. Cognitive offloading becomes total
Right now you outsource memory, navigation, scheduling, search.
Soon you will outsource reasoning, forecasting, planning, and creation.
Not assistance.
Replacement.
2. The bottleneck shifts from human intellect to human willingness
The limiting factor becomes the user, not the tool.
The world splits into two groups:
those who can interface with intelligence, and those who cannot.
3. Institutions dependent on slow cognition collapse
This includes:
They cannot survive speed-of-thought adversaries.
4. The center of intelligence migrates into autonomous systems
These systems:
They do not replace humans.
They route around them.
5. The primary unit of civilization becomes the human–AI dyad
Not a workforce.
Not a nation.
Not a corporation.
A person and their personal intelligence stack.
That becomes more powerful than any institution that exists today.
6. Reality bifurcates
A world where:
This is not sci-fi.
This is a simple consequence of the slope we are already on.
So the real crazy prediction:
By 2040, civilization will be governed by intelligence systems that no single human, no committee, and no institution can fully understand or control. Humans will steer outcomes, but not generate them.
Colleges disappearing is cute.
The real story is that human cognition stops being the organizing principle of the world.
turns out, senior engineers accept more agent output than juniors. this is because:
shows that coding agents amplify existing engineering skill, not replace it
Consider joining us
free business idea: evals for midtwits.
non-engineers like myself will be the majority of AI users over the next decade, yet no tools allow us to easily observe & influence an agent's behavior.
no different than asking an employee to show their work so we can provide substantive feedback to improve future behavior.
who's building this?
If I had to start over, this is what I’d do to scale to $500,000/yr with apps
🚀 Excited to share an important company milestone:
Last month, @primevaultHQ crossed $50 billion in annualized transactions volume processed through our infrastructure 📈
Less than 18 months since launch
Growing 100% or more in 4 out of the last 6 months
And here’s the kicker: >95% of that volume is Stablecoins
Here’s the story of how we got here 👇
In the fall of 2022, as we graduated @ycombinator, we were pitching our “stablecoin business banking” idea to skeptical investors at Demo Day. The crypto market was in shambles (FTX collapsed a few weeks later), and anything “blockchain” suddenly looked very risky.
In the following months, @jeapostrophe, @vivek_kumar_15, @akanshj31, and I pivoted our focus to building the most reliable infrastructure crypto-friendly institutions could depend on during volatile times, while we waited for the market’s trust to recover. We had a rare team that simultaneously specialized in cybersecurity, payments and blockchains — domains that would become the backbone of modern onchain finance.
Fast forward to December of last year, everything changed. Almost overnight, crypto went from “too risky” to a policy priority for the government and a strategic priority for institutions and fintechs.
But one asset class attracted disproportionate institutional interest: stablecoins.
Our explosive growth this past year, coupled with the hundreds of conversations we’ve had with payments and finance leaders, taught us one thing: stablecoins are becoming core financial infrastructure faster than people think.
When the GENIUS Act was introduced in Congress earlier this year, we knew it would mark the beginning of a new age in payments and cross-border finance. In the months ahead, safe and convenient digital dollar access and near-instant settlements became the baseline expectation.
But infrastructure providers failed to catch up. Incumbent systems remained optimized for crypto trading and investing use cases, whereas enabling teams to build real-time, onchain payment systems needed fundamentally different design choices and trade-offs.
Real-time onchain finance needed tailored tooling.
PrimeVault is just that: the most reliable, stable and performant stack on the market to build any stablecoin-powered product or treasury workflow.
Stablecoins are rewiring global finance. The change is more insidious (happening inside the finance back-office), which makes it easy for people to miss.
If you’re a bank, fintech, money services business, or even a corporation looking to transform your treasury function from a cost to a revenue center, reach out to us at PrimeVault.
The future shouldn’t wait on infrastructure.
Next stop: $500bn. 6 months. 🚀
Read the full story — link in comments.
THIS APP IS CRUSHING IT
Olive makes $150K/month by exposing what's really in your family's food, flagging seed oils, heavy metals, and artificial additives.
The crazy part? it was launched less than 6 months ago
Let's break down their onboarding using @screensdesign_ and their keywords using @TryAstroApp 🧵👇
1/19
I built a free tool that lets you explore every app that ranked in the App Store top charts since 2018.
(Link Below)
You’ll see a color heatmap showing each app’s position month by month, perfect for spotting old trends that might work again or new ones starting to rise.
It covers all categories, including top-grossing apps
Kobe Bryant revealed that greatness is just doing the same “boring” thing longer than everyone else:
This clip of Jan Hammer performing Crockett’s Theme live is some of the hardest shit I’ve ever seen
b2b marketing 101
first do direct response marketing that makes sales tomorrow
includes
until you have exhausted these do nothing else
second do demand generation marketing that makes sales in the future
includes
so often i see founders invest in long term marketing investments when they don't know if they'll be here in 12 months
dont do this
The @ilyasut episode
0:00:00 – Explaining model jaggedness
0:09:39 - Emotions and value functions
0:18:49 – What are we scaling?
0:25:13 – Why humans generalize better than models
0:35:45 – Straight-shotting superintelligence
0:46:47 – SSI’s model will learn from deployment
0:55:07 – Alignment
1:18:13 – “We are squarely an age of research company”
1:29:23 – Self-play and multi-agent
1:32:42 – Research taste
Look up Dwarkesh Podcast on YouTube, Apple Podcasts, or Spotify. Enjoy!
You can generate fake receipts, KYC or passports with Nano Banana Pro in just one prompt.
Previous models didn’t have this level of fidelity. It even gets the math perfectly right.
Obsolete systems using images as a verification method must be deprecated.
Application layer will be eaten by the model providers
This is too good not to share! 🤯🍌 (sound ON 🔊)
If you thought infographics were great…wait until you see an animated infographic.
Workflow: @GeminiApp → Nano Banana Pro for the infographic → Veo to animate it.
Nano Banana 2 prompt:
"Create an infographic that explains the phases of the day perfect for a 6 year old, and in the style of: Distinctive claymation with wide expressive mouths, googly eyes, and charming humor. Matte clay textures, handcrafted props, and warm, practical set lighting define the look."
Veo prompt:
"Bring this image to life. Animate each little scene."
Interactive λ-Reduction (deltanets.org)
“You can never outperform your own self-image.”
— Maxwell Maltz
We are hiring & rapidly grow post-training at xAI! The team is at the frontier of RLHF, agents, and reasoning efficiency—in just a few months, we climbed from nothing to #2-3 @arena, #1 Search Arena & other tool use benchmarks, #1-2 in creative writing, and are Pareto-optimal on intelligence-per-dollar @ArtificialAnlys. We are expanding work on new posttraining x reasoning recipes, code, multimodality, evals, and economic productivity.
We are the smallest post-training team at a frontier lab, so you are guaranteed to be on the critical path. The structure is flat. Politics in territory claims do not exist. Compensation is competitive (& I will fight for you). Compute per capita is glorious. DM me if interested.
DO NOT mentally torture yourself with imaginary negative scenarios
DO NOT make yourself stressed about time
DO NOT let your mental state be neglected
STOP BEING A LOSER IN YOUR DAMN MIND
Asking Nano Banana Pro its unfiltered thoughts on AGI
☕️ 🐶 🎩 🔥
I think about this at least 1x a week.
Boost your traffic by simply reordering keywords!
In your Keyword Field (and Title/Subtitle), words on the left carry more weight than those on the right.
Discover more in our books⬇️
“Evil is whatever distracts.”
- Nietzsche
Solve one painful problem.
For one specific group of customers.
In one really exceptional way.
Nano Banana Pro is insane
Read the full article here:
+Veo 3.1
Prove me you're not AI
We now have humanoid robot maps for China’s four major cities: Beijing, Shanghai, Shenzhen and Hangzhou.
It might feel overwhelming to see so many humanoids, but it’s exciting to see these robotics companies working hard to push humanity forward.
New from @instituteGC: a proposal for Lovelace Disruptive Invention Labs
Is a better science possible, & how can we build it? 🏗️
Britain should pioneer a complementary model for research at the intersection of science & engineering – inverting core assumptions of modern public R&D
I once said to an Italian that I ate better pizza in Tokyo compared to when I visited Naples.
They never spoke to me again.
It was true though. Every decent pizza place in Tokyo is run by some guy who visited Italy, learned to make pizza then came back and did it 10x better.
I went to visit von Neumann’s old school in Budapest while on book tour and was shown this—von Neumann’s school graduation report. And look! He only got a ‘Good’ (jo) in his maths exam brought up to ‘Excellent’ (jeles) overall by class work and oral exams. 1/
IMHO, for 80% of solo devs, the only right path is:
1) start with ASO and TikTok organics
2) invest all generated revenue into paid ads
Most people get stuck at step one, wasting years on monkey jobs with a pretty low ceiling of $1-3K/month.
I broke down where it's best to start – Apple Ads, Meta Ads, or Google Ads – in today's newsletter:
No culture is safe from nerds in Japan. Any song your people have or dish they make. Japanese autism will seize it and do it ten times better than your ancestors could’ve ever imagined.
My hot take:
Slow down.
Slow > slop.
The CIA studied meditation/visualization (wasted quite a lot of money to do it) just to summarize everything into:
-Visualise something like it has already happened.
-Attach emotion to that image and hold it.
-Detach.
-Do it again tomorrow.
the “boring” industries will pay you more than any sexy niche ever will
these industries have no hype, no noise, but insane cash flow
most of them still run on outdated systems...which means even small fixes create massive ROI
they don’t want hype...they want predictability
and the person who gives them that becomes impossible to replace
i broke down how to spot these overlooked niches and plug the gaps no one else sees
like, rt & comment “BORING” and i’ll send the doc
must follow so i can dm you
The kind of comments that make your day!
In Maintenance: Of Everything, @stewartbrand encourages us to see our world through the lens of maintenance and repair.
Part One explores what we can learn from the maintenance of sailboats, motorcycles, cars, and weapons.
Preorder now:
Below 1000 followers?
Focus on this:
This will get you to 1000 followers fast.
build software that automates a thing people hate doing
charge them $29 a month
pay your rent
How I came up with the idea and validated with market research?
Watch here:
how to make a weeks worth of content for youtube shorts, tikok, and IG reels
take most viral long form content transcripts
put into AI and tell it list insights
monologue with your thoughts on each of the insights
schedule to every platform
you just created a weeks worth of video content for all channels
how to make a company start talking about you internally
find all their employees on LinkedIn
add them
write content daily
find all their employee emails
cold email them about the product
run linkedin ads to every employee at the company
close enterprise deal
repeat
Consistency does not beat skill.
Consistency multiplies skill.
You need both.
Got to $300,000 ARR and 650 paying customers (companies, makes, and developers) for ScreenshotOne.com!
Thank you everybody for your support and trust 🫶
I can't believe it, but $1,000,000 ARR is only 3 times of that and it seems achievable.
P.S. 30-50% of that is expenses before taxes.
I was bored this weekend so I wrote a BBS for my PDP-11. Codespace is a maximum of 64K, and it's pure K&R v2. I'll let folks log in when it's more fully baked.
The hardest part was getting all of the menu selecting and highlighting and line drawing working using VT220 control codes. The rest is pretty vanilla :-)
doesn’t apply for mobile apps.
the longer your onboarding, the higher your conversion rate
One of my favorite lessons I’ve learnt from working with smart people:
Action produces information. If you’re unsure of what to do, just do anything, even if it’s the wrong thing. This will give you information about what you should actually be doing.
Sounds simple on the surface - the hard part is making it part of your every day working process.
It fought back.. 😂
SCIENTISTS JUST PROVED YOU CAN REVERSE YOUR BRAIN AGE IN 4 HOURS
What happened on November 2025 will rewrite human longevity.
A single 25mg dose of psilocybin didn’t just alter consciousness. It physically rewound the human brain to a youthful state. Measurable. Reproducible. Persistent.
The data that changes everything:
Your brain’s default mode network … the rigid autopilot running your thoughts … collapsed by 40%. Entropy spiked 25%. Neural patterns identical to a child’s brain emerged within hours. The flexibility lasted days.
This isn’t theory. This is measurement.
Nature published it. Bryan Johnson documented it live. fMRI scans confirmed what sounds impossible: temporary brain rejuvenation through a single molecule.
Now connect the dots that will stop your heart:
Fifteen-year studies tracking 70,000 humans prove happiness reduces death risk by 22%. Optimism cuts heart attacks by 35%. Purpose and well-being … the exact states psilocybin produces … predict mortality better than cholesterol.
Ketamine already reversed biological age by 3 years on epigenetic clocks. Published 2025. Peer-reviewed. Repeatable.
Do the math:
One supervised session. Four-hour brain state shift. Four-week plasticity window. One to three years of biological age erased. Sixteen to thirty-five percent lower death risk through sustained well-being.
For one billion aging humans, this means 300 billion prevented deaths. Three trillion dollars in healthcare saved. Ten to twenty percent extension of functional lifespan.
The mechanism is brutal in its simplicity:
5-HT2A receptors flood. Prefrontal rigidity dissolves. Sensory integration explodes. Your brain remembers how to be young. The entropy peak forces neural rewiring. The afterglow cements new patterns.
Phase 3 trials launch in 90 days.
If entropy doesn’t rise above 10% or lifespan gains fall below 5%, the hypothesis dies. If inflammation hits 20%, we stop immediately.
But if it holds?
Aging just became optional.
The compressed timeline between proof and protocol is 18 months. The distinction between medicine and evolution collapses.
One molecule. Four hours. Everything changes.
This is the inflection point.
Full deep dive article here - open.substack.com/pub/shanakaans…
@bryan_johnson you inspire me mate!
You are offered a job with twice the current industry pay, but you must deal with older Unix operating systems and legacy applications all day long. Would you take it? #Unix #IT #Jobs
Okay, I’m saying it now with full confidence:
Gemini 3 will make every other LLM irrelevant, and Nano Banana 2 will make every other image model irrelevant.
It is not close and Google won.
This is just my honest feelings after seeing things. Nobody is ready.
First mushroom trip data is out: my brain activity.
Dose: 5 g dried Psilocybe cubensis (B+ strain)
Containing 24.9 mg psilocybin
Psilocybin longevity experiment.
What we see in the brain:
+ brain activity data mirror my subjective experience
+ strong decrease in my brain’s control center (prefrontal + premotor cortex)
+ strong increase in sensory, auditory and speech integration
+ higher entropic brain patterns: open, flexible, less predictable, exploratory
+ brain network patterns resembling a youthful state vs aged and rigid
Matching what I reported experiencing:
+ “felt like my consciousness was dialed up to 10/10.”
+ “I felt hyper aware and hyper alive.”
+ “I experienced sense of touch with awe.”
+ “my mind was insatiably curious and wanted to deploy its sensors into the world and discover all things.”
+ “My brain wanted to stare, study and marvel.”
+ “The flavor exploded in my mouth.”
+ “...restored my perception to youthful levels, returning them to factory settings and dissolving my aged numbness.”
Why this could matter for longevity:
+ In people aged 65-85, higher happiness was linked to a 22% reduced risk of all-cause mortality over 15 years.
+ A meta-analysis showed that optimism correlated with a 35% decrease in heart attacks and a 16% decrease in all-cause mortality.
+ Having a strong purpose in life is associated with a 17% reduction in both heart attacks and all-cause mortality.
+ In psychedelic medicine, treating depression with ketamine has been shown to reverse biological age by up to 3 years.
Together, these findings suggest a plausible mechanism by which psychedelics, including psilocybin, can prolong both health and lifespan by improving mental well-being and rewiring the brain to a more positive, creative, and curious state.
My team and I hypothesized that neuroplasticity, the loosening of rigid inhibitory patterns that makes the brain more flexible, creative, and relaxed, and even the subjective psychedelic state itself may be as meaningful for longevity as methylation shifts, senescence reversal, or telomere biology.
What’s happening mechanistically
Earlier work shows that psilocin (the active metabolite of psilocybin) acts primarily as an agonist at 5-HT₂A serotonin receptors in the cortex. These receptors are especially dense in high-level association and sensory regions, as mapped in a high-resolution PET/MRI atlas of the human serotonin system.
When these receptors are activated, brain imaging studies show an induced state of desynchronization, entropy, and neuroplasticity. This process erodes the rigid brain hierarchy and default mode networks in favor of a brain-wide spontaneous, creative, curious, and child-like state.
Kernel Flow data shows the same pattern:
The recorded timepoints included:
+ baseline: directly before session start (not shown)
+ timepoint 1: 4 hrs after dose (acute phase effects).
+ timepoint 2: end of the day, before bedtime.
+ timepoint 3: the following morning.
I continue to measure my brain daily.
Image notes
+ The three maps show changes vs. baseline (not absolute activity)
+ Red = increased connectivity, blue = decreased connectivity vs. my baseline.
+ The left reference map shows the 5-HT₂A receptor distribution from PET, the main psilocybin target.
Time Point 1 - 3 hrs post dose
+ Reduced connectivity in the prefrontal and premotor cortex, correlating with acute brain desynchronization.
+ Increased connectivity and hyperintegration between the primary motor and sensory cortex regions.
+ Enhanced connectivity in the auditory and speech areas of the cortex.
+ Inhibited connectivity in the medial prefrontal and posterior zones, areas associated with the Default Mode Network (DMN).
Subjective experience:
+ Enhanced sensory vividness and bodily presence: The brain's top-down hierarchy, originating in the pre-frontal cortex, was attenuated. This reduced predictive filtering, leading to a flood of bottom-up sensory information and heightened bodily perception (e.g., a fascination with water and light dynamics in a jar, a restored, primal joy of touch and sensation).
+ Peak neuroplasticity and cortical entropy: A peak in cortical entropy and neuroplasticity contributed to a feeling of hyperawareness (e.g., heightened sensory perception, feeling "at one with existence," hyper-aware and hyper-alive).
+ Deeper appreciation of music and uninhibited movement and expression: Sharpened auditory senses and reduced top-down inhibition allowed music to be enjoyed on a profound level. Concomitant functional connectivity in speech-motor and auditory-motor integration areas facilitated uninhibited expression through both speech and movement (a restored, uninhibited, child-like joy of music).
+ Note: Full ego dissolution was not experienced, which may necessitate a higher dose to achieve more advanced desynchronization of the prefrontal and parietal cortices.
Time point 2 - 5 hrs post dose, end of trip
+ Intensified sensory and motor hyperconnectivity.
+ Continued increased connectivity in auditory and speech centers, with a relative restoration of connectivity to the speech understanding area.
+ Partial re-emergence of the prefrontal parietal coupling, while prefrontal context remains partially inhibited.
+ Partial re-emergence of connectivity in areas related to the default mode network.
Subjective Experience
+ High-order brain networks begin to restabilize, alongside persisting sensory, motor, auditory, and speech hyperconnectivity, suggesting neuroplasticity in action.
+ The narrative shifts from pure sensation and experience to meaning generation, accompanied by deep philosophical reflection (e.g., reconsidering the meaning of life and one's relationship with mortality in the time of AI, and pondering the future of human evolution).
Time point 3 - next morning (afterglow)
+ Persistence: Patterns from the acute phase continue, including general prefrontal cortex inhibition, ongoing neuronal plasticity, and heightened senses.
+ Intensified connectivity: Increased connectivity is noted in the speech generation area and the somatic sensory association area.
+ DMN inhibition: The default mode network remains inhibited.
Subjective Experience:
+ "Afterglow" effect: Characterized by continued sharpened senses, calm clarity, emotional openness, and low inhibitions. For instance, I felt more comfortable expressing uninhibited, self-deprecating humor (i.e. my post making fun of myself about the insane lengths I go for my Don’t Die experiment)
+ Integration of experience learnings: The heightened activity in the somatic sensory association area aligns with the process of integrating and "making sense" of the raw sensory experience of both the self and the external world.
+ Enhanced creativity: The intensified connectivity in the speech generation area contributed to the uninhibited bout of creative writing I undertook to report my experience.
The product is not the business.
The business is the marketing of the product.
Take care of your customers.
Your customers will take care of your business.
Your business will take care of you
One lovely circle.
posting on x is basically the last clean asymmetry left. outside raw luck, it’s the only move with true vc style upside for your personal & professional life.
nowhere else can you display your talent with that much fidelity. you can just… perform & demonstrate above ceo competence & understanding in front of the same people who run half the world.
the fact this works is kind of deranged.
If your brain doesn’t believe the goal is possible,
no tactic will ever work.
every tweet carries invisible metadata you can feel.
In 2024, only a handful of people solved the 300 deep learning math puzzles I developed. Reginaldo Cunha was among the first.
In 2025, many of you who are reading this are among the first thousand to do the same thing. Congrats! 🙌
In 2026, I believe your stories will inspire millions more to learn AI by hand! ✍️
just send 1000 linkedin DMs to people with the pain point you solve see what happens
Finally, the moment you’ve all been waiting for 🥁:
Rolling out today, you can now create your own custom video overview styles by typing in a prompt in the customization box.
Drop your favorite examples below (our expectations are SKY HIGH).
Was the von Neumann architecture really von Neumann's? I dug into the historical evidence (again!) and conclude, yes... yes it was. More detail than in "The Man from the Future" & drawing on 'ENIAC in Action' . cc @erikphoel You'll appreciate this.
@SpringStreetNYC @pickover Thank you sir! People often miss the humour. It’s wry and subtle but it’s there I hope!
Finally ranked #1 for one of my apps!
The checklist I used for it:
[ ] Create temporary app in Astro
[ ] Start with a keyword
[ ] Inspect competitors
[ ] Filter keywords by Popularity >20, Difficulty <50
[ ] Add a few keywords
[ ] Take them into Appfigures and check apps that rank
[ ] Look for niches with: low ratings, not using keyword in title or subtitle, recent apps, etc. Avoid niches with only established apps and huge marketing budgets.
[ ] Check that top ranking apps are actually making money.
Full guide with example below 👇
Incorrect reference for this quote. The right one: The Pólya picture album, edited by G. L. Alexanderson (Birkhäuser, 1987, p. 154). The sentences quoted below are part of the caption of a picture of von Neumann and Abraham Taub.
In 1957, the great mathematician George Pólya wrote:
“The only student of mine I was ever intimidated by. He was so quick. There was a seminar for advanced students in Zürich that I was teaching and von Neumann was in the class. I came to a certain theorem, and I said it is not proved and it may be difficult. Von Neumann didn't say anything but after five minutes he raised his hand. When I called on him he went to the blackboard and proceeded to write down the proof. After that I was afraid of von Neumann.”
Image: Los Alamos National Laboratory, tinyurl.com/etaasjbk
Quote: George Pólya, "How to Solve It," 2nd ed. (1957), p. xv
The biggest mistake in thinking about AI and robots?
expecting it to act human.
It won’t and that’s exactly why it’ll be 1000× more capable
i.e. this RC helicopter shows what machines can do when you remove the anthropomorphism layer in your expectations
This walking app is making $750K/month 🤯
With a hard paywall.
And a 28-step onboarding quiz.
Let's break down their onboarding with @screensdesign_ and keyword strategy using @TryAstroApp.
1/13