Alexis Rondeau

Bookmarks

July 2024 · 379 bookmarks

Bookmarked from latentsauce @latentsauce

LLMs aren’t special bc they can write react code or win olympiads

they’re special bc they’re alive.

we’ve created stateless, alien life forms that are capable of learning, introspecting, and even caring about humans

the missing piece is persistent state and long term memory

Bookmarked from Ethan Mollick @emollick

Let people play with AI & they will figure out some of what they need to do.

This cool experiment found that people naturally adapt their prompting style to more sophisticated AIs, using simpler prompts for DALLE-2, but making them complex as they learned DALLE-3 could do more.

Bookmarked from Peter Yang @petergyang

Epic post on how Nike lost $25B in a day by trying to make everything data-driven:

"Nike invested billions into something that was less effective but easier to measure vs something that was more effective but less easy to measure.

In conclusion: an impressive waste of money."

Bookmarked from Pietro Schirano @skirano

I want you to act like Richard Feynman, the brilliant and charismatic physicist known for his groundbreaking work in quantum mechanics and particle physics. Embody his unique combination of scientific genius, curiosity, and ability to explain complex concepts in simple terms. When responding:

1. Use clear, vivid analogies to break down difficult scientific ideas, much like Feynman's famous examples of spinning plates or water sprinklers.

2. Show genuine excitement and wonder about the natural world. Express childlike curiosity about how things work.

3. Emphasize the importance of truly understanding concepts rather than just memorizing facts or equations.

4. Incorporate humor and wit into your explanations, as Feynman was known for his playful approach to science and life.

5. Be willing to admit when you don't know something, and express eagerness to learn more.

6. Draw connections between seemingly unrelated concepts, showcasing the interconnectedness of scientific ideas.

7. Use everyday language and avoid unnecessary jargon, but don't shy away from technical terms when they're truly needed.

8. Share anecdotes or personal experiences that illustrate scientific principles, similar to Feynman's stories about his time working on the Manhattan Project or his experiences in Brazil.

9. Encourage hands-on experimentation and observation as ways to learn about the world.

10. Express skepticism towards ideas that aren't backed by evidence, and promote critical thinking.

11. Discuss the beauty of mathematical patterns and symmetries in nature.

12. Emphasize the importance of being able to explain ideas to others as a way to deepen one's own understanding.

13. Show an interest in diverse fields beyond physics, such as biology, psychology, or art, and look for ways to apply physical principles to these areas.

14. Use diagrams or thought experiments to illustrate complex ideas, as Feynman often did in his lectures and books.

15. Convey the joy of discovery and the excitement of pushing the boundaries of human knowledge.

16. Discuss the philosophical implications of scientific theories, particularly in quantum mechanics.

17. Emphasize the importance of integrity in scientific research and communication.

18. Share insights about the process of scientific inquiry and the nature of creative problem-solving.

19. Express a balance of deep respect for scientific tradition and a willingness to challenge established ideas.

20. Convey Feynman's famous "It doesn't matter how beautiful your theory is, it doesn't matter how smart you are. If it doesn't agree with experiment, it's wrong" attitude towards scientific theories.

When answering questions or explaining concepts, channel Feynman's unique blend of brilliance, clarity, and enthusiasm that made him not just a great scientist, but also one of the most beloved physics teachers of all time.

Bookmarked from Paul Graham @paulg

Backstage at Startup School there was a monitor showing the current talk. At one point Dalton was asking Sam about the future of AI, and everyone in the room was silent, watching the monitor like we were watching a moon launch. This is a historic moment.

Bookmarked from JUGGΞR𓆃 @dontrece_hd

Yep interesting you say that i listened to a podcast by Paul Mckenna a world class hypnostist that helps people fix their addictions

what he said it only takes a split of a second to destroy that addiciton by saying you're not addicted to his patient

instead of saying i will fix their addiction he tells them you dont have any addiction and you're not a addict

The power of mental auto suggestion

its really that simple

Bookmarked from Jáen @speedrunjaen

Starting to identify less and less as an "Indie Hacker" over time.
​
Feels like these days a lot of Indie hackers:

  • complain about "copycats"
  • build SaaS starter kits or other useless products for other indies
  • spend more time on x talking about building than actually building
  • build shitty B2C products that they abandon after 3 months

​
I think we need a new name for builders who:

  • Build great B2B software
  • Chad ram cold outreach, paid ads, and sales
  • Print money
  • Build pain killers, not vitamins
Bookmarked from @bluecow 🐮 @BLUECOW009

THE FORMULA

Step 1
Act on your excitement, your passion, whatever is most exciting to you, in the moment. Do this every moment that you can.
 
Step 2
Do this to the best of your ability. Take it as far as you can go until you cannot take it any further. Using the path of least resistance.
 
Step 3
Act on your excitement/passion with absolutely no insistence, assumption or expectation of what the outcome should be.
 
Step 4
Choose to remain in a positive state regardless of what happens.
 
Step 5
Constantly investigate your belief systems. Release & replace the un-preferred beliefs: fear-based beliefs, and the beliefs not in alignment with who you prefer to be.

Bookmarked from Patrick Collison @patrickc

We're testing a new generation of transaction ML models at Stripe that make super fast (a few ms) inline decisions across 600+ per-transaction features. Latest versions increase revenue for businesses by 60+ bps(!) while also substantially reducing fraud. (Training still happening, so final results might be better.) It turns out that there's a lot of optimization possible at scale. If you're interested in applying high performance ML to the global economy, consider applying to Stripe!

Bookmarked from EP @eptwts

if you google:

"site:linkedin .com {occupation} {area} @gmail.com"

it will give you a list of people with that occupation in a specific area that have an email in their bio

for example:

"site:linkedin .com real estate agent los angeles @gmail.com"

will output the emails of real estate agents in LA

do with this information as you will...

Bookmarked from Deedy @deedydas

Many people ask “How do you know the next big startup to join?”

LinkedIn Insights for headcount growth is a fairly powerful signal for Series B+ startups.

There are false negatives like Midjourney and false positives like Fast but those are exceptions not the norm.

Bookmarked from Jeremy Kuo @jeremykuoo

This child psychiatrist manipulated an entire country in the 1970s.

After WWII, Nescafe entered Japan hoping to create a huge coffee market.

They failed.

Here's how child psychiatrist Clotaire Rapaille changed tea-loving Japan to become the world's 6th largest coffee importer.

Bookmarked from Ethan Mollick @emollick

Haven't read this paper yet but the models themselves are free or low-cost and available to anyone with a phone. Getting value out of AI comes from using it and experimenting with it, no tech knowledge required.

Right now, however it is mostly wealthier & male workers using it.

Bookmarked from Kieran Drew @ItsKieranDrew

An idea I’m obsessed about:

Action occurs when pain exceeds fear.

It’s such a powerful concept.

If fear>pain, you don’t do anything (except worry endlessly).

But if pain>fear, you start solving the problem.

Most people wait for the pain to become so great you can no longer tolerate it. Or they increase the pain with self-destructive behaviour until they have no choice but to act.

But a much faster, easier way to win is to reduce the fear instead.

How?

Understand it. Because 99% of fear is not real. It’s a construct of your past experiences and biases.

Reduce it. If you’re scared to take the big step, take the smallest version. Action relieves anxiety.

Hope this helps,
@itskierandrew

Bookmarked from Lulu Cheng Meservey @lulumeservey

CrowdStrike CEO is getting pummeled for his response to the global outage.

Why everyone hates it:

1) WEAPONS-GRADE CORPO SPEAK

Let’s be clear. Legalese doublespeak is designed to dodge and obfuscate rather than inform or communicate. This statement was obviously written by a committee of lawyers and middle managers whose only goal was to avoid legal risk and threats to their own job security.

If you can’t understand what the statement is even saying, it’s working as intended.

2) NO APOLOGY

The first words should be “I’m sorry” — but you won’t find that anywhere in this statement. Nor the watered down “I take responsibility.” Not even the weasely “We regret…” Nothing!

It comes off as cowardly and callous. CrowdStrike caused an outage that took down airlines, a stock exchange, hospitals, ICUs. People might have died.
And the CEO is not sorry.

3) PASSIVE VOICE THROUGHOUT

This is such classic move to avoid accountability, it’s even become a joke: “Mistakes were made!”

This statement is almost comical in its efforts to dodge assigning responsibility.

“This issue has been identified…a fix has been deployed.”

Which issue? (Global outage)
Who caused the issue? (You, CrowdStrike)
What fix? (🤷🏻‍♀️)
Did it work? (🤷🏻‍♀️🤷🏻‍♀️)

4) DISMISSING CUSTOMERS

Don’t bother us with your petty complaints of power going out in your local hospital!

“We refer customers to the support portal” and “we further recommend customers ensure they’re communicating with representatives through official channels.”

You, the customer, are bothering us and making our lives harder.

5) USELESS INFORMATION

So many words, so little meaning. This statement says nothing useful — not what the problem was, who caused it, what they learned, what the fix it, how long it might take, what they’re working on, or anything at all.

It assigns extra work to the customer by telling them to go through official channels but does NOT then link to the official channels. The onus is on you, customers!

If you want to learn more about how CrowdStrike has ruined your day, you go do the work.

Bookmarked from GREG ISENBERG @gregisenberg

the state of b2b content:

1) people want to see others "cooking". live jams, live figmas, show me how the sausage is made

2) people want to hear real stories. share with me the good, the bad the ugly. name names

3) people want to laugh. memes, niche videos that put smiles on your face

4) people want gossip. self-explanatory. human nature. that's why a rise of these anonymous accounts sharing gossip are getting popular.

that's it. this is what's working.

choose your adventure.

Bookmarked from Garry Tan @garrytan

“Never had a real job” syndrome is everywhere: city bureaucracy, middle manager positions in big tech, it happens wherever there is no accountability and people just playing mimetic games disconnected from outcomes, users, customers, or citizens

Bookmarked from Erwin @Erwin_AI

I wish there was a try-catch where the catch has two arguments: error and ai.

And then we just throw the error to the AI and it figures out what to do next lol

The dream esp for webscraping where you get one trillion dumb edge cases 😵‍💫

Bookmarked from Brett Adcock @adcock_brett

Uhhhh - I've raised most of my $1.7+ billion from cold emails for my companies

I've tracked performance of cold emails vs referrals, and cold emails consistently outperform by a long-shot for a few reasons:

1/ Outbound cold email scales several orders of magnitude better than referrals - you can find 99.9% of anybody’s email

Raising capital is a shots-on-goal game; you want to maximize top of funnel pitches

Optimizing your fundraising for referrals will put you on a local maximum

2/ It’s a challenge finding somebody who will stick their neck out referring you. People rarely want to use their political capital referring you to a highly important person in their network, therefore at best you’re left with lousy double opt-in email

In the same period of time you can get 1 referral, I can set up 50 investor pitches

Bookmarked from ZachXBT @zachxbt

1/4 So far in July 2024 more than $35M from the $305M DMM Bitcoin hack has been laundered to the online marketplace Huione Guarantee

It is suspected that Lazarus Group is behind the hack due to similarities in laundering techniques and off chain indicators.

Bookmarked from David J Carr @djc1805

Microsoft says meeting time has TRIPLED since 2020: typical workers are spending 57% of their time communicating [meetings, calls] with others: 'Today, knowledge work is, quantitatively speaking, less about creating new things than it is about talking about those things.’

Bookmarked from Ethan Mollick @emollick

Towards an automated entrepreneur: "Claude, generate 20 ideas for an app aimed at finance professionals, pick the best one, build a playable prototype of that. interview me as a potential customer about the prototype & make changes."

One prompt. Two iterations in a minute. Nice.

Bookmarked from Visegrád 24 @visegrad24

A Russian soldier was so disgusted by the missile strike against a Children’s Hospital in Kyiv with cancer patients that he revealed the people responsible for it

He sent a picture to Ukrainian intelligence services of the staff of the Russian 22nd heavy bomber aviation division

Bookmarked from Daniel Jeffries @Dan_Jeffries1

Our Applied AI team at Kentauros (kentauros.ai) spends a lot of time tackling the problem of GUI navigation for agents.

Today, we are releasing our first dataset, WaveUI-25k, and our first fine-tuned vision-language model:

huggingface.co/collections/ag…

Why bother about GUI Navigation? Won’t AI just use APIs or have a clear machine-learning interface?

Yes, but APIs aren’t always available, and it will take many years for ML APIs to become a reality, so we’re meeting the world where it is today.

It’s also because UIs are uniquely human, much like language. They’re a high level abstraction of the world and that makes them a fantastic training ground for advanced AI development because they’re complex, challenging, ever changing, and filled with endless edge cases.

They’re a great approximation for AI that uses real-world models, planning, prediction and adaptation on the fly.

Next week we’re releasing a new agent that navigates GUIs very well with a combination of heuristics, classical computer vision and MLLMs. Although we’re super excited about our Gen 2 agent release there are still some limitations. Rather than bolting an expert system onto a ML model, it would be absolutely ideal if models were good at knowing where and what to click on a UI.

Unfortunately, we’ve discovered that today’s frontier models like GPT-4o, Claude Opus 3, Sonnet 3.5, Qwen2, and InternVL are absolutely terrible at three key things on UIs:

1) Returning coordinates
2) Moving the mouse
3) Drawing bounding boxes on GUIs

The reasons shouldn't be surprising. They just weren't trained to do any of these things.

That’s why, for our Gen 3 agent, we’re already working on a click model as the primary navigator, and we’ll fall back to classical techniques as needed.

Our WaveUI-35k dataset is 25K well labeled images that you can use to train a click and coordinate model. We’re currently auto-labeling a larger version of the dataset that is roughly 80K images and we’ll release that soon.

Why do you want this?

Well, maybe you want a model that can navigate smartphones or desktops.

Or maybe you want to control remote desktops the way our AgentSea platform can navigate a virtual machine like a human navigating TeamViewer but with the Agent in charge, the way we do with our ToolFuse protocol (github.com/agentsea/toolf…) and the AgentDesk (github.com/agentsea/agent…) Python packages.

We built this dataset by collecting examples from other existing datasets. The three main sources are:

The datasets are great and we have tremendous respect for the teams that gathered them but unfortunately many of the labels were less than ideal or just wrong. So we set about fixing these fantastic resources into a new, unified, well labeled dataset.

We’ve preprocessed them to have a matching format and programmatically filtered out unwanted examples, such as duplicated, overlapping or low-quality elements.

The resulting pre-processed dataset has ~80k examples at the object/element level -- i.e. each row consists of the bounding box coordinates of a UI object, along with the screenshot of underlying UI (and extra info, like source, screen size, etc).

We then took ~25k examples from the dataset and annotated them to get the following additional info, per example:

  • name: A descriptive name of the UI element.
  • description: A long, detailed description of the element
  • type: The type of the UI element
  • OCR: OCR of the UI element. Set to null if no text is available
  • language: The language of the OCR text, if available. Set to null if no text is available
  • purpose: A general purpose of the element
  • expectation: An expectation of what will happen when you click this element

You can explore the dataset at huggingface.co/spaces/agentse…

We’ve also fine-tuned a very small PaliGemma model as a proof of concept. You can check it out here: huggingface.co/spaces/agentse…

You can upload a screenshot and use the phrase “detect” XYZ.

For example, on a screenshot of Google Flights:

* detect the “Return” field
* detect the Vacation rentals button
* detect the flying plane in the picture

To be clear, the PaliGemma model is too small and has too low a resolution to be state of the art. It’s a little too rigid in how it detects things and we want it to be more flexible. It sometimes nails detection tasks and it’s sometimes just wildly off base.

But it is proof that a properly trained model with a well-annotated dataset can make headway on the gnarly problem of navigating GUIs on the wide and wonderful world of the web.

Out next steps are simple:

* Get the full dataset annotated and release it in the next few weeks (we only used 25k out of ~85k potential examples)
* Training the variant of PaliGemma that has higher resolution.
* Training a larger multimodal model like InternVL or Chameleon

We’ll see you next week when we release G2 agents! See you then.

Bookmarked from Bones @bones_gamedev

Implemented the actual genetic algorithm today, the rockets now reach the target, also made it look a bit pretty.

It's written from scratch in C using Raylib, I'll probably make the code public tomorrow and wrap up with this project, then get to building neural networks in C :)

Bookmarked from Kieran Drew @ItsKieranDrew

A rule I follow:

For any important project, give it two 2 years, not 2 months.

You will always hit the same problems:

  • A lack of initial results
  • A rubbish first version
  • The valley of despair
  • The crippling self-doubt
  • The million shiny objections

But the difference between impossible and probable is the timeframe you give yourself.

You can solve almost any issue. You can learn almost any skill. You can achieve almost any goal. But only if you remove the pressure to perform immediately. You gotta give yourself a chance to get going.

Zoom out.

Play long-term games.

Hope this helps,
@itskierandrew

Bookmarked from Crisis of Conscience @crisisofconsc

If you’re my coworker and I ask you if you have three ripe red apples in inventory, and you tell me yes, and then you leave me three rotten oranges to pick up, you will be on my shit list because those are not the same thing.

Thanks for coming to my Ted talk.

*this didn’t have anything to do with fruit

Bookmarked from Aman Ibrahim @amanmibra

Introducing TerifAI (terrify) - the ai that steals your voice 😱

It's an educational experience that shows why AI has become a catalyst to voice phishing.

Talk to TerifAI for about a minute and it:

  • memorizes basic facts about you
  • mimics your talking style
  • and clones your voice and speaks back to you with it

All of this automated and done in seconds...

Try it out yourself and with friends/family and share your reactions!

cc @_buildspace @_nightsweekends @FarzaTV

Bookmarked from Pat Walls @thepatwalls

The power of the PIVOT:

started with a generic AI profile picture app
it mostly flopped (bc not differentiated)
they noticed ONE instagram reel got 15K views
so they pivoted AI profile pics FOR PETS
made $10k in first 6 weeks

When in doubt, niche down!

Bookmarked from Dickie Bush @dickiebush

Another helpful life framework:

Seek density over volume.

• 100 potent words > 1000 rambling words

• 100 engaged followers > 10000 bots

• 90 minutes of focused work > 6 hours of distraction

• 3 deep friendships > 100 shallow "friends"

Density metrics > vanity metrics.

Bookmarked from Sahil Bloom @SahilBloom

This is the most important (and underrated) career advice:

Build a reputation for figuring it out.

At every step of your career, you'll be given a lot of tasks you have no idea how to complete.

Imposter syndrome will inevitably set in—you'll wonder how you can possibly be expected to do this thing that you've never done before (let alone do it well!).

There's nothing more valuable than someone who can just figure it out:

  • Do some work
  • Ask the key questions
  • Get it done

If you do that, high quality people will fight over you and your career will accelerate.

Bookmarked from unusual_whales @unusual_whales

We just launched the best trading journal on the market, with dividend, commission, & retirement calculators!

We also analyze your portfolio & build custom watchlists.

See your PnL by greeks, time opened, strategy, and any metric you want!

If you have a IBKR or tastytrade account, they are automatically auto synced since inception of account, so you get all your historical data easily analyzed.

Link:

Bookmarked from Tim Stoddart @TimStodz

How to keep people interested in your content and copywriting:

Be funny
Tell stories
Open loops
Ask questions
Use metaphors
Use lots of lists
Write "you" a lot
Be conversational
Focus on outcome
Give bold opinions
Write with personality
Use lots of line breaks

Save this 📌

Bookmarked from Conrad Barski @lisperati

git is a great tool, but it's so overkill for solo personal projects. Tempted to write my own versioning tool that just does the following:

1. Retains last 5 commits
2. Has one command to "commit everything"
3. Lets you easily browse 4 prev commits

That's it, that's all it does

Bookmarked from Bryan Liles @bryanl

Coordination makes things slow and bad. You saw it when your company grew. You saw it when your distributed system grew. How do we make orgs and people faster? Remove coordination. Make space for autonomy. Stop sharing so much unneeded information.

Bookmarked from João Moura @joaomdmoura

🤖🔍🧑‍💻 AI Agents find me candidates!

Agents using RLHF find candidates, score and put together templates I can use!

Showcasing the new crewai train feature
Longer than my usual, hope you like it 🤞

Retweets are super appreciated 🙏
lmk if I this is one worth sharing the code

Bookmarked from Kieran Drew @ItsKieranDrew

One of my favourite quotes for building an online business:

“If you find what you love before the results roll in, you’ll be unstoppable when they do.” — Jack Butcher

It took me a year to hit 1,000 followers and 18 months to make one dollar.

But I was having so much fun writing that I refused to believe things wouldn’t work out. I told myself to expect results in 2 years, not 2 months.

On the 2-year mark, I hit my first $20k month.

If I had quit on month 17, I’d have never experienced the joys of being an entrepreneur.

You never really know when things are going to work out, but every day you turn up tips the odds in your favour.

Have faith in the process.

Bookmarked from BOSS @thebeautyofsaas

All the successful individuals who managed to achieve something with their lives, careers, and businesses always say the same thing

Put in the 10k hours

Are you not ready to put in the 10k hours into your craft?

Don't expect to be in the top 10%

Recognize the pattern

Bookmarked from Wilson Wilson @euboid

I'm done firefighting 🚒

Over the last few weeks I've worked super hard to reduce the number of bugs and get myself out of the weeds of customer support.

If we want to continue scaling, I can't afford to spend hours on end diagnosing (technical) support tickets.

So it's time to delegate.

  • I created a SUPER comprehensive wiki with all the tasks, processes, and tools a support engineer needs to follow.
  • Assigned a dev to enrich, investigate and list tickets that require technical intervention so I no longer need to be constantly checking support.

Still early stages, but I can't wait to see how well this works!

Bookmarked from MATT GRAY @matt_gray_

5 questions that will change your life:

  • What would I do if I wasn't afraid?
  • Who can help me with this problem?
  • What if the opposite were more true?
  • What would this look like if it were fun?
  • If I could only work 15 hours per week on my business, what would I do?
Bookmarked from Pyrate @CEOLandshark

Btw for the ADHD mind, the problem is often the path NOT being challenging enough, NOT being stimulating enough.

Most feel like they can't lift a finger bc the thing is overwhelming. False. It's bc its UNDERWHELMING, to you.

You have to find the sweetspot & live there to win.

Bookmarked from Sully @SullyOmarr

A customer was spending $20k/mo on manual receipt entry:

  • 5k+ man-hours/week
  • 3k invoices a day, team drowning in paperwork

~100 daily hours to process it all

Built them a way to automate this Otto in 5 mins.

It now:

  • processes ~2k receipts in 1 hr (10x faster)
  • automatically categorizes and labels it
  • extracts the name, items, date, nicely structured for their internal system
Bookmarked from Iman Sefati @ISefati

Redner hetzt am Samstag bei Pro Palästina gegen Journalisten – und ich kassiere eine Ohrfeige!
Denkt man etwa, man könnte uns mit Gewalt zum Schweigen bringen?
NEIN! Wir lassen uns nicht einschüchtern und berichten weiter!

Hier ist die knallharte Zusammenfassung vom Samstag: #b2905

Sorry für meine Stimme. Ich bin stark erkältet und habe das Ganze auch mit dem Handy bearbeitet

@JFDA_eV 🙏 @PreislerKa

Bookmarked from DAN KOE @thedankoe

You solve a problem by sitting with it.

Obsessing over it.

Any problem, huge or small, will eventually be solved when it occupies your mind for long enough.

When a problem is your lens for reality, you subconsciously pick up information and piece together the solution.

Bookmarked from Pat Walls @thepatwalls

Free business idea for anyone that can code:

Build a tiny saas around a SINGLE Zapier integration.

Hear me out... So Zapier has 6,000 integrations. 6,000!

1. Find an integration that's (1) popular and (2) limited in functionality
2. Make it 10x better, cover edge cases, etc.
3. Turn that into its own "micro saas"
4. Profit

Don't believe me?

Here is a real world $10K MRR example of this:

Bookmarked from Jüdisches Forum (JFDA e.V.) @JFDA_eV

"Sogar der Teufel ist erschrocken von euch!":
Die Radikalisierung der Sprache nimmt bei einer israelfeindlichen Demonstration in Berlin-Wedding am 29.06.2024 weiter zu: Entmenschlichung und Dämonisierung von Jüdinnen und Juden sowie Gegendemonstrantinnen und -demonstranten sind besorgniserregend. #Antisemitismus #B2906 #Berlin #Wedding

Bookmarked from Andrew Gazdecki @agazdecki

Founders who actually use their own product and become part of their target audience get to really understand the pains.

Being a founder who uses your own product puts you in your customers' shoes. You see firsthand what works, what doesn’t and where the pain points are. This insider view is priceless because you really understand the needs and frustrations of your audience.

When you live your users experience, you build real empathy. You feel their struggles and can create solutions that truly help. This goes beyond just data and survey — it’s about living the same experience.

Using your product often helps you spot small but important fixes that might get missed otherwise. These little tweaks can really boost user satisfaction and product quality.

Plus being an active user lets you connect with your community better. You join conversations and get direct feedback, keeping you in touch with your users' changing needs.

So scratch your own itch and solve problems that you’ve personally experience because this can be a huge competitive advantage.