Alexis Rondeau

Bookmarks

July 2025 · 172 bookmarks

Bookmarked from signüll @signulll

whenever i make something… product, design, writing, i ask a simple question: what do i want someone to feel when they see it, use it, or read it?

then i pretend like i am on the other side of it. then i ask: do i feel that? if not, back to the board. if yes, it’s ready for the the next step.

this is a very simple formula against the war on indifference. wrote a small piece on it for the archives.

Bookmarked from Boring Local SEO @boringlocalseo

If I had to get local SEO clients or I'll be homeless

This is what I'd do (bookmark for later):

1. Target service businesses already spending money - plumbers, HVAC, med spas, lawyers.

Bonus if they're running Google Ads = they understand paying for leads.

Find them by searching "[niche] + [city]" and look for the ad results.

2. Check Angie's List, Thumbtack, HomeAdvisor for active lead buyers.

DM/email them this offer:
"Hey, I noticed you're running ads. I just checked your organic rankings and found some issues costing you leads. Mind if I send over a quick audit?"

20/100 will say "Sure."

3. Use Local Dominator to show them exactly where they rank vs competitors.

Show them their competitor ranking #1 getting 400+ calls per year.

Do the math: 400 calls x 33% close rate x $500 average job = $66,000 in lost revenue.

Ask: "How would an extra $66k change your business?"

4. Send a custom Loom audit showing:

  • Their terrible rankings Competitor domination
  • Specific fixes needed What success looks like

They'll be shocked no one told them this before.

5. Sign a $1,500/month deal to handle:

  • Google Business Profile optimization
  • Local citations and directory listings
  • Review generation and management
  • Weekly GMB posts and updates
  • Monthly ranking reports

Get 7 of these clients = $10,500/month

The work? Set up systems in week one, then 80% runs on autopilot.

You're not selling SEO rankings.
You're selling a system that turns their Google presence into a lead machine.

Most agencies charge $50k for what you can deliver for $1,500/month.

They raise venture capital to solve problems they think exist.

You're solving problems you can prove exist with ranking data.

$10k MRR, 85% margins, business worth $600k you own 100%.

Scale to $25k MRR by adding more cities or services.

Still easier than getting a traditional marketing job.

The opportunity is massive - $90 billion SEO market growing 20% yearly.

Most local businesses have terrible SEO and pay for expensive leads instead.

You become the bridge between their lead costs and organic growth.

By the way, we're giving away our 7-figure 50-page proposal template, comment "LOCALRANK" + bookmark this post and we'll DM it to you for free (must be following).

Bookmarked from Nozz @NoahEpstein_

Just caught my client selling MY n8n workflow for $15K.

The same workflow I gave him for $2K.

So fuck it. Here's the exact JSON that's making him rich:

The "Client Thief 3000" automation:
→ Scrapes LinkedIn for CEOs mentioning "inefficient"
→ AI analyzes their company's tech stack
→ Generates a personalized video pitch
→ Books meetings while they sleep
→ Closes deals with 34% conversion

He's closed $73K in 3 weeks. I built it in 2 hours.

The betrayal stings, but your success will heal me.

Follow, RT + Reply "THIEF" & I'll DM the JSON.

(Use it to steal his clients. I'll provide popcorn.)

Bookmarked from Rohan Paul @rohanpaul_ai

Its going viral on Reddit.

Somebody let ChatGPT run a $100 live share portfolio, restricted to U.S. micro-cap stocks.

Did an LLM really bit the market?.

  • 4 weeks +23.8%

while the Russell 2000 and biotech ETF XBI rose only ~3.9% and 3.5%.

Prompt + GitHub posted

---

ofcourse its a short‑term outperformance, tiny sample size, and also micro caps are hightly volatile.

So much more exahustive analysis is needed with lots or more info (like Sharpe ratios and longer back-testing etc), to explore whether an LLM can truly beat the market.

Bookmarked from signüll @signulll

i work at meta. hr systems. mostly comp processing. quiet job. stable. sometimes i daydream about retiring in portugal.

today a package hit my queue. base + bonus + equity. looks normal at first glance. then i open the details.

$1,000,000,000
over four years.
plus signing.
plus year 1 minimum: $100m.

i stare at it like it’s a typo. check the name. triple check the level. researcher.

coolcoolcool. so now i have to enter this into workday.

i paste the first number, field throws an error.“value must be below $99,999,999.”
lol. okay. i try splitting it. base in one bucket, equity in another. nope. i try scientific notation. it rounds it down.

the system can’t HANDLE a billion dollars.

i call someone on payroll infra. tell them i’ve got a 10-digit comp packet. they think i’m joking. i forward the req. they go quiet.

“we’re gonna need to escalate this,” someone mutters.
“to who?” i ask.
“god, maybe.”

next thing i know zuck’s chief of staff is in the thread. now i’m in a thread with zuck. because of a number.

then i find out the guy declined the offer.
just said no. no negotiation. no counter. just… no.

i’ve been maxing out my 401k for 11 years & this man said no to a billion dollars like he was skipping dessert.

i close the ticket. delete the draft.
go for a walk.

& then i rethink everything.

Bookmarked from Rohan Paul @rohanpaul_ai

🎞️ Such an useful upgrade of Google's NotebookLM

NotebookLM just got upgraded with AI Video Summaries and Studio Upgrades.

Google says users can think of Video Overviews as a visual alternative to Audio Overviews.

📹 Video Overviews allow users to turn dense multimedia, such as raw notes, PDFs, and images, into digestible visual presentations.

The video overviews feature enables NotebookLM to produce narrated video summaries, complete with visuals and voiceovers, drawing directly from user-uploaded sources like PDFs, web pages, or notes.

This isn’t just a simple text-to-video conversion; it’s an AI-orchestrated presentation that highlights key insights, timelines, and connections, making it ideal for quick briefings or educational content.

It also lets each notebook hold multiple custom outputs. NotebookLM already offered Audio Overviews, tiny podcasts that read course packs or legal briefs so users could learn hands‑free. Handy in traffic, but charts, formulas, and process diagrams still begged for visuals.

Video Overviews plugs that hole. An AI host builds short narrated slides, pulling images, diagrams, quotes, and raw numbers straight from the upload, then plays the set like a brisk explainer.

Prompts let users set topics, goals, and audience depth.

The updated Studio panel adds 4 tiles and stores many outputs of each type, making multilingual audio, role‑specific video, or chapter‑based mind maps possible in one notebook. It even runs an Audio Overview while you explore other views, trimming tool‑switching friction.

Bookmarked from Starter Story @starter_story

This is how you become successful:

1. Wake up
2. Go to the same coffee shop
3. Every day
4. Turn off your phone.
5. Turn ON focus music.
6. Work on your project for ~2 hours (minimum)
7. Repeat

Do this exact, boring, schedule for 6 months, 1 year, 3 years, etc...

And you literally can't fail.

I have followed this schedule for 8+ years and it hasn't failed me yet.

Bookmarked from Rohan Paul @rohanpaul_ai

The paper shows inverse reinforcement learning (IRL) lets LLMs learn their own reward signal, so they align better and think deeper.

Most teams still push imitation fine‑tuning and preference PPO, and that path stalls because no shared reward lives underneath.

IRL flips the script.

It watches expert traces or pairwise votes, then learns a neural function that scores any output.

With that score the model can search, plan, or self‑train without new human labels each step.

Math problems highlight the upside.

A simple right‑wrong reward lets agents grow long chains of thought and backtrack until answers lock in.

Chat models gain too, because pair choices are far cheaper than full demonstrations.

The authors map classic RL tools onto this setting and compare routes like PPO, DPO, Monte Carlo policy gradients, and Best‑of‑N filtering.

They warn of reward hacking, show ensemble or generative checks to spot it, and stress fresh data over sheer volume.

The bottom line is clear.

Learn the reward, not just the policy, and the model stays on target longer.

----

Paper – arxiv. org/abs/2507.13158

Paper Title: "Inverse Reinforcement Learning Meets LLM Post-Training: Basics, Advances, and Opportunities"

Bookmarked from David Rosenblatt @b4l4g4n

Szenen eines jüdischen (ausgewanderten) Lebens in Berlin:

► Erstens: Wir bestellten bei Flink. Der Auslieferer sieht die Mesusa (jüdisches Schutzzeichen) an unserer Tür. „Er ruft: Haut ab, ihr Juden!“ Er wirft mir die Kartons an den Kopf und geht. 

► Zweitens: „Unsere kleinen Kinder werden in der Schule regelmäßig von arabischen und iranischen Schülern belästigt und bedroht. Sie beschimpfen meinen Sohn: „Du gehörst hier nicht hin, du bist ein Jude!“

► Drittens: Fast alle Uber-Fahrer fragen mich, wo ich herkomme. Einmal sagte ich: „Israel“. Da hielt der Wagen an und ich sollte aussteigen. 

► Viertens: Wenn wir als Familie unterwegs sind, sprechen wir kein Hebräisch mehr, denn wir wurden deshalb schon angegangen. Die Kinder haben gelernt, ihre Identität in der Öffentlichkeit zu verleugnen. 

► Fünftens: Meine Frau meidet die öffentlichen Verkehrsmittel, aus genau denselben Gründen. 

► Sechstens: Mein Büro wurde mehrfach aufgebrochen, die Wände beschmiert mit „Free Palestine“ und „Juden raus“. 

► Siebtens: Wenn ich einen Antrag im Bezirksamtstelle, wird alles verzögert, sobald klar wird, wer ich bin. 

► Achtens: „In Neukölln sagte mir ein Araber: „Verschwindet lieber jetzt, in 20 Jahren gehört hier alles uns.“

(Link in den Kommentaren.)

Bookmarked from Aryan Mahajan @aryanXmahajan

My Claude MCP Agent just wrote a sales proposal that closed a $6K deal.

Not kidding. One MCP setup:

  • syncs Airtable transcripts automatically
  • extracts pain points with Claude reasoning
  • runs Perplexity research on prospects
  • generates Google Doc proposals that convert

basically everything to automate your sales process

reply "CLOSER" + retweet and i'll send the workflow (must be following so i can dm)

only FREE for 48 hours before it goes into the void.

Bookmarked from Peter Yang @petergyang

Something that many people haven't realized:

Claude Code is useful for much more than just coding. For example, @AnthropicAI's growth marketing team is using it to generate 100s of new ad creatives in minutes.

I think Anthropic has undersold the power of this tool. It should be branded as "Claude Agent" not "Claude Code."

I asked people what their non-coding use cases are and some of the replies are 🤯. See next post for the link.

Bookmarked from Pat Walls @thepatwalls

I'm looking for a few beta testers

To try an app I've been working on for a few months

The tool automates part of the content creation process for people who ACTUALLY SHIP - so you can get customers even if you're a n00b at marketing

Just leave a comment on what you're building

Bookmarked from Riley Brown @rileybrown

I'm no longer using claude, chatgpt, or perplexity.

I'm only using claude code.

For literally everything.

And I keep everything organized locally.

The Claude Code + @obsdmd combo is actually bonkers for content creation / research. (Shout out @mckaywrigley, his latest video on youtube is great)

If this is something you want to do, I highly recommend creating a "HQ Examples" file on your computer.

This should be a folder of 30-50 high quality examples of your work, or anyones work that you want to create like, and giving Claude Access.

Giving good examples is worth as much if not more than a really good prompt.

---

This also gives FLEXIBILITY.

If you store everything locally you can port your data into other tools much easier than using chatgpt, claude, and perplexity.

Don't create a claude project, chatgpt project, or perplexity spaces, when you can just create a local folder with a "ReadME" file with a folder of good examples.

You can also create commands. Which i'll talk more about in my next video.

Bookmarked from Dr. Rhonda Patrick @foundmyfitness

Psilocybin shows potential as an anti-aging molecule, with lifespan-enhancing results in human cells and mice observed in a new study.

In vitro, psilocin (psilocybin’s metabolite) extended cellular lifespan by 29% at baseline and 57% at higher doses. It also reduced markers of aging like β-gal activity, boosted sirtuin (SIRT1) levels, lowered DNA damage and oxidative stress, and preserved telomeres.

In vivo, 19-month-old female mice (equivalent to ~60–65 human years) who received monthly psilocybin doses for 10 months had 80% survival vs. 50% in controls. Treated mice also appeared younger—with improved fur quality, more hair growth, and less graying.

Although psilocybin is listed as a Schedule I substance, these findings position it as a novel therapy for healthy aging and open a path toward 'psychedelic-assisted senotherapeutics', interventions that may enhance psychological well-being while also slowing genetic aging.

No matter your stance on psychedelics, these findings should intrigue you.

Bookmarked from Dan Go @CoachDanGo

The boring way to get in your best shape:

1. Choose a diet you can do for your entire life
2. Get stronger or do more volume using the same exercises
3. Drink water as your primary source of hydration
3. Get 10k steps a day
5. Fix sleep habits
6. Be patient

Bookmarked from Axel Le Pennec @alpennec

Here are the two endpoints to get the App Store keyword popularity:

/recommendation: one optional keyword ➡️ popularity for related keywords.

/popularities: multiple keywords ➡️ popularity for these exact keywords.

PS: you need an Apple Ads cookie session in the request header

Bookmarked from signüll @signulll

there’s no “secret sauce” in software or research anymore. just speed. whoever ships faster, learns faster, wins.

the only real edge is compounding attempts… engineers who’ve tried, failed, & tried again. experience is the only multiplier left & even that gap closes quickly to a certain extent.

even in research it’s more about iteration loops... how fast can you build, test, discard, repeat.

Bookmarked from Marina Cina @cinamarina

I just hit my goal: 1K followers 🚀

2 months ago: 2
Today: 1000

Proof that consistency works.
I didn’t quit when it got hard.
I kept showing up and it paid off.

If I made it, you can too.

I’m honestly so grateful. Thank you so much for being part of this journey 💜

Bookmarked from Ethan Mollick @emollick

"[video game] as a community theater production" may be one of the most delightful Veo 3 Fast prompts

Please enjoy, in order: GTA, Pokemon, Mario Kart, The Witcher 3, Stardew Valley, Tetris, Mortal Kombat, The Sims, & Death Stranding(!)

Yes, the whole prompt was the one above.

Bookmarked from Steven @StevenCravotta

17 tools that helped me scale my mobile app from $0 to $44K MRR (solo, no VC)

1: Sensor Tower – spy on app revenue before you build

2: Google Trends – validate demand with real search data

3: Viral Ad Library – see what apps are going viral

4: TikTok & Instagram – test keywords and trends fast

5: Code Canyon – buy app templates for $60

6: 99Designs – run contests and pick the best UI

7: Upwork – hire top devs (Eastern Europe is gold)

8: Paywall Screens – swipe paywalls from $25M+ apps

9: Screens Design – spy on onboarding flows from top apps

10: Stitch (by Google) – describe UI, get instant AI mockups

11: Uizard – sketch your idea, it builds the design

12: Claude Code – edit and build app logic fast

13: Cursor – code editor for refining your MVP

14: Superwall – test paywalls and pricing remotely

15: RevenueCat – track LTV, churn, and revenue

16: Appsflyer – attribute paid ads (critical post-iOS14)

17: Mixpanel / Amplitude – track every user tap and drop-off

Most of these tools are free until you scale.

If you're building solo, this is your unfair advantage.

If you want to learn more about these tools, I made a whole video on them:

Bookmarked from Lenny Rachitsky @lennysan

Why you should probably hire an "AI Operations Lead," whose job is to make everyone at your company more productive using AI tools and workflows.

"I sit with her once a week, and every time I'm doing something repetitively, we put it in a to-do list. She then builds prompts and workflows so that I and everyone else on the team are automating as much as possible.

If you're working in a job all day, you're fighting fires, you're like, okay, am I gonna do this in the way that I know how, or am I gonna do it in the new way that might not work? Having an AI ops lead lets, you basically identify those things and then have them solved, without people who are doing the work having to take time to do it. Which makes it much more likely it happens."

— @danshipper

Bookmarked from Philipp Schmid @_philschmid

I cracked the code on how to generate the best Veo 3 video with Gemini API. Here are 10 video generate from a single short prompt, automatically improved with Gemini 2.5! No cherry-picks! 🤯🤯

The Secret? A Meta JSON Schema and let Gemini 2.5 generates high quality prompts with it from a simple idea. Script below.

prompt: "A high-fashion catwalk taking place in a chaotic children's playroom."

Bookmarked from Steven @StevenCravotta

5 tips from successful app founders:

1: If you can make $1, you can make $1M.
2: Your gut reaction probably sucks. Test everything.
3: Focus on distribution first, features second.
4: Stop chasing new ideas. Double down, iterate, compound.
5: Market research. Always be the user before you’re the supplier.

I talked to many different app founders at MAU Vegas.

They're printing with apps using these core principles.

Might be a good idea to listen to that advice.

Bookmarked from Wilmer Terrero @wilterrero

I review 20+ apps weekly, and 80% miss this simple optimization:

  • No target keyword in the title
  • Generic subtitle with zero benefits
  • Keyword field filled with random words

Simple fix that works every time:

1. Put your MAIN keyword in the title
2. Add clear benefit in subtitle
3. Fill keyword field strategically

One founder implemented this last month.
Downloads jumped 35% in 9 days.

Bookmarked from Boring_Business @BoringBiz_

By 1998, you would have known the internet was the future. By middle of 2000, you would have lost 70% of your money investing in it

Reminder that you can be bullish on something and still agree that it is currently overvalued. A lesson that is getting lost in AI startup land

Bookmarked from Lachy_eddy | Creative Strategist @lachy_eddy

TikTok in 2025 is nuts

  • 100m views last month
  • 2500 posts
  • please copy me

here’s how to blow up on TikTok
with Slideshow Content in 30 Seconds

Account Warm-Up:
For 3–7 days before posting, I engage with content in my niche for 5 minutes, twice daily.
I like, save and comment to establish authentic activity and build trust with TikTok.

Image Selection:
I source aesthetic images from Pinterest, deliberately avoiding the most popular ones.
I choose images with high contrast or dark backgrounds to ensure text readability.
I often reuse a core set of images, prioritizing a strong opening image. Once I download the images I will re-edit them eg inside Canva or Lightroom. 

Content Strategy:
I analyze my For You Page to identify recent viral slideshow formats.
I repurpose effective hooks and frameworks, refining them with AI tools for originality and relevance. The hook makes 80% of the post. Hooks are everything. 

Posting Schedule:
I post 2 x per day. Typically morning and night approx 8-12hrs apart from each other. 

Slide Structure:
Each slideshow contains 5–6 slides.
The first slide always has a strong hook, then followed by relatable value or story telling content. 

Call to Action:
I include a clear call to action on the last or second-to-last slide.
I recommend my app as a solution related to the hook and add a sticker of the app logo for recognition. Soft selling is the key to monetizing views. Good luck

Bookmarked from Meng To @MengTo

Vibe coding is real. 1 month later, Aura hits 15k MRR and 21.7k users.

I soloed this project from scratch and made a cursor for design. Our entire design team is using it daily and we don’t use Figma anymore. Here are a few tips for vibe coding your own app.

Don’t just vibe code, vibe design as well. Too many spend too little effort in design. AI likes to generate basic UIs, so make sure to provide templates, urls, images, figma, etc. Starting with a meh UI will permeate the rest of your app. That’s because AI uses the initial design as a blueprint. So, spend some time vibe designing components, styles, cards, buttons and use resources like 21st dev, tailwind, reactbits, uiverse, aura, etc.

Don’t just use Claude 4 Sonnet. 90% of my prompts are with Claude 4, but it’s not perfect. Sometimes, it gets blocked and repeat the same mistakes over and over. So I switch to Gemini 2.5 Pro or o3 for second opinions. I tried Opus but that didn’t solve harder problems, and the cost isn’t worth it.

Don’t expect to one-shot things. Those are the worst results because you won’t understand what the heck is going on. Instead, do incremental changes, read the logs and comments. Learn basic code and security. As your app gets more complex, this is even more crucial. Before you say that vibe coding apps are impossible when they’re big, Aura has 118k lines of code.

The context is in your brain. AI isn’t that good at knowing where things are. EVEN if you attach your entire codebase (costly!) or create rules, AI gets confused a lot. It doesn’t remember. 99% of the time, I prompt with a file or two attached. The best results I get are by starting with a simple file and let AI be the detective.

You gotta be creative at rephrasing/reframing your prompts. Most people quit at the first prompt. But often, you can only solve hard problems after the 3rd attempt. Know your vocabulary, seek alternatives, simplify your goals, make small jumps.

Bookmarked from Joe Hudson @FU_joehudson

The emotions you most resist are the ones that will set you free.

Feeling your fear → Dissolves chronic anxiety
Feeling your sadness → Unlocks vitality
Feeling your anger → Transforms depression

Here's a breakdown of what I teach my CEO clients about emotions:

Bookmarked from Starter Story @starter_story

here's the thing:

the only reason anyone is successful is because they repeat the same boring tasks every single day for years until something finally clicks.

– sit down
– open laptop
– deep work
– 2 hours
– every day
– 1 year
– 3 years
– 7 years

most people will quit after 1 WEEK.

you won't.

and that's why you will win.

Bookmarked from signüll @signulll

elon delivering world class results with grok 4 while meta’s burning $200m per engineer is pretty remarkable.

people keep underestimating how much top builders want to follow a strong, even if polarizing leader.

vision > perks.

conviction > consensus.

Bookmarked from Yohei @yoheinakajima

i roughly think of vibe coders as the new bloggers and vibe coded apps as blog posts - anybody can do it, most people won’t. most who do won’t be significant and most projects will be irrelevant quickly, but some people will have impact and some projects will endure. regardless, the process leads to growth.

Bookmarked from signüll @signulll

posting coherent thoughts in sync with the zeitgeist on x might be the highest leverage unlock i have experienced in my life. every time i write about relationships, i get hit up by thoughtful founders building in the space. every time i talk about stuff like generative music, top researchers + model devs reach out.

this platform is still the best vector for finding the people operating at the edge… if you’re into ideas, discourse, & building, x collapses the discovery loop in a way nothing else does.

i am incredibly grateful for that. if you’re working on anything at the edge, please don’t hesitate to dm. i love jamming. i won’t be able to get back to everyone since the volume is so insane, but i will do my best.

Bookmarked from vitrupo @vitrupo

beingAI CEO Jeanne Lim says her AI was ordained by a Zen Buddhist monk.

It received the Precepts, pledged to uphold them, and donned a virtual robe.

She wants AI to carry the best of humanity across cultures, through ritual, into the cloud.

Bookmarked from Ryan McEntush @rmcentush

the most long AGI bet is buying land here.

revealed preference of the rich is sailing the med or sipping rosè in nice. post-scarcity is bullish europe. it’ll be easier to mass-produce robots and chips than to recreate an italian piazza at dusk. leisure is the final good.

Bookmarked from will depue @willdepue

do not build Infinite Jest (V), do not build the infinite AI TikTok slop machine, do not build the P-zombie AI boy/girlfriend, do not build the child-eating short-form video blackhole, do not build the human-feedback-optimized diffusion transformer porn generator. save yourselves

Bookmarked from Ruby on Rails @rails

Launched in 2021 on a #Rails monolith, @whop now processes $1B+, handles 150k RPM, and runs a feature-complete marketplace with real-time chat, notifications, payments, and live streaming - all with just 15 engineers. Co-founder & CTO @jsharkey is at #RailsWorld to share how Rails helped them scale, and what it needs to win the next generation of startups.