Riley

ai gets you maybe 40% of the way to something good. the other 60% is you caring enough to keep going. this rule holds for prompting, but also bigger stuff like building sites and pitch decks. one shotting and light engagement produces something, but it's crappy. steering, questioning, fine-tuning, knowing what to ask, over and over gets you something worthwhile.

testing ai tools without a project keeps you shallow. saw this with myself - i've dabbled with image and video gen, but didn't go deep until i needed them for my site. ended up spending days on mood boards, prompts, animating site frames, directing transitions and music. having a real need made me so much better at both. start a project and watch yourself seriously grow.

i work with ai all day and can't imagine doing 80+% of my tasks without it. i straight up wouldnt work for someone not taking advantage of these capabilities. companies leaning in will cut workers, but companies leaning out will lose them too.

a question that's getting hard to answer: did you do it or did ai do it? we both did...?

ai is so much better when you know what to ask. learn what to ask by following people in domains that you're trying to grow into. as a vibe coder, i'm tuning in to people talking about security and architecture. there's no way i could prompt my way into some of the tips they give. so much value taking what you find outside of the chat and bringing it in for ai to evaluate and integrate.

@KennyAlami that one’s a little deceptive. doesn’t seem like its worth building custom, im planning to switch to riverside or screenpal since loom wasn’t meeting my needs. did cancel kit though and they were smart not to email me and make it into this screenshot 😅

@chasing_next How'd you replace Loom? I need to cancel my n8n too lol

yesterday wasnt a good day for saas...

anyone have experience transferring usernames between accounts on here? i want a new one, but dont want to lose the one i have.

finding the skill that helps most with ai is the same one that helped me in business: seeing how one decision impacts 5 other things. since ai moves faster and lets you do more, understanding connections is increasingly important. to be a good ai manager, you need to be able to forecast, track, and act on ripple effects at an even greater scale.

@_ItsmeJJ_9 Sounds cool!

@chasing_next Haha, I just built a dashboard for my event business. If a dummy like me can do it, there’s hope for everyone!

you can build anything now, that's not the hard part anymore. the hard part is the strategy, finding an audience, and selling to them.

@MartinUushona @kilocode 👀

we're in a thought bubble. the everyday user isn't thinking about ai. they're thinking about their role and what's on their plate. if a tool helps with things like invoice processing, lead routing, compliance checks... that's where the win is. most workers don't have time to figure out how to integrate ai themselves, figure it out for them. huge inroads for ai solutions that are industry or vertical-specific beyond tech.

openai is just like us.

@amirmxt ha totally, now i just need to meet nancy.

@chasing_next Its nancy in hr & compliance

i have yet to meet a copilot super user.

markdown is the holy grail for ai right now, but i don't think it's forever. works great for a solo operator building a second brain, but convincing a team to use .md over current visual interfaces is rough. for compounding ai to really take off it needs to be approachable to teams. this means plugging in to the tools they already use (sheets, docs, excel, word, ppt...). not siloed connectors, but converting info behind the scenes so ai has context across format types, without downgrading the everyday user's experience. i think this piece is coming sooner than later.

a year ago you needed to know how to code at least a little bit to create a quality site. today you don't have to know at all.

have had quite a few people who use chatgpt or gemini ask me about claude. my take is that unless you or your team is proficient at the tool you have, you shouldn't think about switching. swapping one tool for another wont magically get you to use ai beyond the basics. that said, if you're already competent and systemizing processes with ai... then there could be something to gain from switching.

if you're building a product, highly recommend creating an ai customer panel for on demand feedback. here's how i did this: - transcribed customer discovery calls with tactiq - analyzed calls to strengthen icp profile and who they interact with - had claude create team of icp reviewers i call on demand my audit team includes the main buyer, their boss, various end users, and the budget approver. i point them at a page or idea and they each tell me what's confusing, missing, or doesn't land for their role. they come up with ideas and issues i'd never think of.