Jason Cohen

56.8K posts

Jason Cohen banner
Jason Cohen

Jason Cohen

@asmartbear

Keyword, buzzword, half-truth, adjective, hey look at me! (𝘧𝘰𝘶𝘯𝘥𝘦𝘳 𝘰𝘧 𝘵𝘸𝘰 𝘶𝘯𝘪𝘤𝘰𝘳𝘯𝘴: https://t.co/Cc4OvZx0T9, https://t.co/JTEGCe7Zq6)

Join 65,000 subscribers → Katılım Aralık 2008
713 Takip Edilen46.6K Takipçiler
Sabitlenmiş Tweet
Jason Cohen
Jason Cohen@asmartbear·
List everyone you know who started by launching a startup every month, and the ninth one was really successful. List everyone who has 1 really good idea / year, and obsessed over it, and was successful. The second is 100x more common. We can all list the people in the first category, exactly because there’s so few of them. Whereas the latter describes every single unicorn startup, as well as nearly all bootstrappers. So many you can’t even name a small fraction. This is because startups are hard, good ideas are few, there needs to be good reasons for it to work, they are not haphazard A/B tests, they take 6-24 months to really get going even in the success case, they require obsession and dedication. It’s like thinking you’re going to become a concert pianist by getting 5% through a piece and then moving on to the next one, 20 times. That doesn’t add up to 100%. That means you’re probably pretty good at the 5%, and a shit musician, and unsuccessful. Doing 10 reps of failed startups gives you “reps” in setting up default infrastructure and having AI make a Next.js template with auth. That is not a “rep” at building products that people want to buy. Doing 10 reps of failed “marketing” where people aren’t attracted, and the ones who are attracted bounce off, and the ones who don’t bounce off don’t buy, and of the 12 that ever bought, 8 cancelled, is not “learning.” You still have no idea how to market or sell products that people want to buy. Being at a small, successful startup for 1 year would teach you 100x more than all those “reps” at failure. You 𝘤𝘢𝘯 learn from failure, but you learn 10x more from success. Failure doesn’t even teach you “what not to do;” some people do that same stuff and succeed. Now list all the people on this website who have launched 7 ideas they knew were weak from the start, thinking they will be the next @levelsio, where their products are all doing $670 in MRR, just a string of failures, getting no closer to “success.” Don’t try to mimic outliers. Outliers are very interesting, exactly because they’re rare and uncopyable. So why are you trying to copy the strategy that works 1/100th as often? Getting “reps” at failure and calling it “learning?” And before you say “What about 𝘺𝘰𝘶! Why should we copy you!”, remember that for nearly two decades I have been saying “do not copy me.” I’ve written about rampant survivor bias in business advice including my own, and my “path” to PMF calls 𝘮𝘺𝘴𝘦𝘭𝘧 out for not necessarily walking that path. You shouldn’t copy anyone. You should cherry-pick things you like -- attitudes, frameworks, tools, insights, prioritization schemes, decision philosophy -- and further change them to match yourself. That’s good; you should accelerate yourself using others’ wisdom. That’s not a “copy.” Try to find a good idea that’s actually worth doing. Get obsessed and stay obsessed. Get inspired by others, cherry-picking things that make you a better version of yourself, but be yourself. Stop building 1% of a bad idea and calling it “learning.”
English
70
59
634
114.8K
Jason Cohen
Jason Cohen@asmartbear·
If you have 1000 servers, each experiencing a fatal error randomly once every 3 years… you will get random, unpredictable, unstoppable failures every single day. Scale is hard because it makes rare things common. longform.asmartbear.com/scale-rare/?ut…
English
0
2
8
869
Jason Cohen
Jason Cohen@asmartbear·
Q: What if the intersection of what I like, what I’m good at, and what the world needs, is currently the empty set? That's admirably wise, honest, and introspective of you. You should be proud of being that -- almost no one can (or at least will) be those things. It means you're in "explore" mode instead of "execute." You know that the primary challenge is to find the right project, as opposed to executing a specific project. This, in turn, suggests the right sort of actions or goals. Actions include exposing yourself to more potential opportunities. That might be working in a field, or pointed meetups, or interacting in communities where potential co-founders of potential projects might be lurking, or experimenting with ideas of your own, building with intent-to-prototype-and-test rather than intent-to-scale, or consuming certain books / articles / podcasts related to ideation or discovery. All that said, of course there cannot be a formula that produces great ideas in 30 days, though there are "courses" that purport to, and perhaps some of them are worthwhile, not because they can deliver on that promise literally, but because they can increase the odds of you finding the right idea, and sooner. Still, working on the wrong thing is definitely a waste of time, so while exploration is frustrating (because you'd rather be focussed on the right project), hopefully you can also see it as fun and exciting, since of course it 𝘪𝘴 that as well. More on finding yourself: longform.asmartbear.com/pivot-points/?… More on Explore vs Execute: longform.asmartbear.com/explore-execut… More on finding fulfillment: longform.asmartbear.com/fulfillment/?u…
English
0
1
11
716
Jason Cohen retweetledi
Jason Cohen
Jason Cohen@asmartbear·
𝗣𝗲𝗿𝘀𝗶𝘀𝘁𝗲𝗻𝗰𝗲 𝗛𝗮𝗶𝗸𝘂 When you can’t take it, Go for a walk and touch grass. Back to it, renewed.
English
2
2
16
870
Jason Cohen
Jason Cohen@asmartbear·
@jojo_data Exactly. They're already wishing your product will be the answer.
English
0
0
1
27
JoJo 🔨 LegalCents
JoJo 🔨 LegalCents@jojo_data·
@asmartbear This changed how I approach demos. I stopped trying to create urgency and started asking "what does your current process look like?" The moment they describe the pain themselves, the product sells itself.
English
1
0
0
45
Jason Cohen
Jason Cohen@asmartbear·
The sales pitch is showing the customer that this product fits their needs and constraints. If you find yourself convincing them they need to buy something, or that their needs are stronger and more acute than they thought, that’s a waste of time.
English
3
1
18
1.7K
Harish Krishnan
Harish Krishnan@Hk05Krishnan·
@asmartbear Reverse-chronological sorting for a top-down document is a classic case of Engineer Logic colliding with User Reality.
English
1
0
1
27
Jason Cohen
Jason Cohen@asmartbear·
So annoying how Google Doc’s “comment list” pane is in reverse-order by date, which means if someone leaves comments from top to bottom -- as one nearly always does, the comment list is in exactly the wrong order. At least give us a “sort by” toggle?
English
3
0
8
1.7K
SiteWarming
SiteWarming@SiteWarming·
That's something I've seen in a lot of salespeople.. they seem to confuse persuasion with qualification... If the need isn't already there, you're creating urgency, and customers can sense that. That's why it's better to simply target the people who are already looking for your solution, or else you'll end up wasting time and resources...
English
1
0
1
60
Jason Cohen
Jason Cohen@asmartbear·
It is true of all successful startups, that “the founder never gave up.” So it becomes a “law of success.” Of course, sometimes people don't give up, but never find success. So, it's necessary, but not sufficient. longform.asmartbear.com/chaos-at-start…
English
3
1
23
1.7K
Jason Cohen
Jason Cohen@asmartbear·
@jacklhanlon While they just 𝘵𝘩𝘪𝘯𝘬 it's a toy, because indeed the raw chat is pretty bad, they will not realize what it 𝘤𝘰𝘶𝘭𝘥 be, and it's a self-fulfilling prophesy. For now anyway.
English
0
0
0
14
Jack
Jack@jacklhanlon·
@asmartbear Perhaps I shouldn't have given explicit numbers. The sentiment is that there is a "minimum quality threshold" where the AI tool goes from being a Toy in a companies experimental budget, to a critical product in their workflow
English
1
0
0
15
Jason Cohen
Jason Cohen@asmartbear·
No one cares whether it was "made with AI," but many people are interested in trying things that allow 𝘵𝘩𝘦𝘮 to use AI. However the key word is "try." Cancellation can be 90% at 4 months. It has to also work, which it typically doesn't (currently).
English
12
0
43
2.9K
Jason Cohen
Jason Cohen@asmartbear·
@jacklhanlon No AI is 97% accurate. ChatGPT is maybe 60% on simple questions. Claude code is much better with context and honed skills. But not 87%.
English
1
0
0
51
Jack
Jack@jacklhanlon·
@asmartbear Great advice I heard recently. An AI product with 80% accuracy that saves 10% of the time is a Toy. An AI product with 97% accuracy that saves 10% of the time is a business. AI gave us flexibility, workflows require consistency.
English
1
0
0
50
Jason Cohen
Jason Cohen@asmartbear·
If a fast decision is wrong, that’s more OK, because you find out fast, and change it fast. Slow decisions, the opposite. If “slow” dramatically decreases risk of it being wrong, make it slow. Else, make it fast. More fast/slow tips: longform.asmartbear.com/decisions-fast…
English
2
1
20
1.3K
Jason Cohen
Jason Cohen@asmartbear·
The first iPhone was (ironically) terrible at being a phone, but people overlooked all the problems because the design and the “real internet” was so magical. What’s your thing, so revolutionary that customers will overlook all objections? longform.asmartbear.com/the-important-…
English
2
3
15
1.5K
TechGray CA
TechGray CA@ca_techiegray·
@asmartbear Most outages I debug start with 'it can't be my code.' Humility is the only five-nine error-handling.
English
1
0
1
11
Engineer
Engineer@SpeedTestDemon·
@asmartbear not entirely -- an important overlooked half of Icarus's story was not to fly too low to the sea else the dampness could make the wings too heavy to flap. the story is really about golden mean.
English
1
0
0
11
TadasG 💻
TadasG 💻@tadasgedgaudas·
@asmartbear I don't think we have anymore products like the first iphone today. Maybe EV cars like Tesla/Chineese EVs?
English
1
0
1
37