David

@doodles doodle me!

experience unlimited creativity. makedoodles.ai

@AlexReibman I would like this as well

Does anyone have a good agent skillfile that: - Spins up a screen recorder - Runs your app end to end - Edits and annotates the video to show all the core highlights of the feature working Plus if not limited to webapps/GUIs either

@MitcheIl launch

@outboundman Or they forget to start new sessions

@dzhng context usage is the killer one. most people burn the window on setup and then wonder why the outputs get worse halfway through

It's very easy to be net unproductive with LLMs LLMs have a "zone of genius" that's a function of: - agent friendliness of environment (e.g. clean code) - task complexity - current context usage When it stays in its zone of genius, it's great and can one shot tasks perfectly However things degrade VERY quickly to a slop fest the moment you get outside of its zone of genius As an AI coder your job is to ride the fine line of maximizing LLM output while keeping it in its zone of genius at all times. Once it's outside of the zone, no amount of promoting is going to save you Productivity only comes when you use your boost when you need it, while avoiding the potholes along the way If you blindly use LLMs for everything, it's going to net out at zero

I actually partially agree with this - llms are great coders but terrible architects. The problem is most people don't know the difference and either blindly trust the llm or don't give it a chance.

@nzieber @transitive_bs

Where is the AI hub type venue in Bangkok?

@andrewchen Or another way to put it is cowork / openclaw class of products are the antithesis of vertical saas

the magic of cowork and openclaw and other AI products is that they replace our giant row of infinite browser tabs And lol - no, don't feel guilty, I have too many tabs too. AI makes it so that every workflow that required 4 browser tabs and a spreadsheet is getting collapsed into one AI-native experience Just as one quick example- think about how you used to research a person or a company: LinkedIn tab, X tab, Google tab, notes doc, slack open. now one prompt does it in 10 seconds. the "tab count" of a workflow is basically a proxy for how much AI can compress it if your product eliminates 6 tabs and a copy-paste loop, users will like it. If you can create a whole series of these workflows then your users will absolutely love it. Thus the biggest opportunities are workflows where people currently alt-tab 20+ times per task. Sales, recruiting, research, compliance, procurement. Boring? yes. Massive? also yes. But this is why these agentic tools are going to crush AI doesn't need to be superintelligent to be wildly useful. it just needs to be good enough to close the tabs

@C_Ghillie @the_judge1111 bingo

@the_judge1111 @dzhng probably at the $800M round a few years ago. $9B/$800M = ~11x * 75% (25% dilution on funding rounds) * 80% (20% carry) = 6.6x gets you pretty close.

6.3x on my one and only angel investment, should I be doing more 😂

@the_judge1111 You're taking my post too literally 😂

@dzhng Angel investment??? That you’ve 6.3x’d??? With the company priced at 9 stacks? My brother that was a Series C Praying you understand the economics of the SPV you’re in

@_lilyzhng @modal @crcolgrove Congrats!

Shout out to @modal for hosting this great panel at NVIDIA GTC, I'm excited to join @crcolgrove and Robin next Monday! We'll be talking about 𝘴𝘦𝘳𝘷𝘦𝘳𝘭𝘦𝘴𝘴 𝘎𝘗𝘜, 𝘢𝘨𝘦𝘯𝘵 𝘰𝘳𝘤𝘩𝘦𝘴𝘵𝘳𝘢𝘵𝘪𝘰𝘯, 𝘢𝘨𝘦𝘯𝘵𝘪𝘤 𝘦𝘷𝘢𝘭, and 𝘧𝘶𝘯 𝘱𝘳𝘰𝘥𝘶𝘤𝘵𝘴 𝘸𝘦 𝘢𝘳𝘦 𝘣𝘶𝘪𝘭𝘥𝘪𝘯𝘨. Everyone talks about models, but but who knows what keeps us up at night😅. Come hang if you're at GTC! luma.com/niwjk62h

Random thought - since code is now a commodity and trending towards being 100% AI generated, does having so many layers of abstraction actually make sense? Example: we use Expo for @duetchat - it's great most of the time but we have a ton of long tail bugs that is blocked by Expo. After a certain point maybe it's easier to just delete Expo and go direct to native iOS and Android instead. The write-once-run-everywhere value prop becomes diluted once you realize that AI can just automate test every single view on every single platform to ensure consistent behavior.

@andrewchen The goal is to always have enough modularization & test coverage where we just accept (prob 70-80% there, mobile expo app is biggest blocker for new) We've already sectioned off parts of our codebase where humans won't be allowed to touch in the near future

One question I've been asking founders is: do you try to review all the code that the LLMs write or do you just accept it? I think it's about 50-50 right now but the momentum is towards just accepting the AI-generated code and I think that number will eventually go to 100% This is one of the most telling indications of how AI-native a team is. It's hard to get super high throughput if you are reviewing every line Poll: what do you do?

@thekitze What's something that the lobster does that you think would be hard to replicate for other agents?

in theory this sounds great but i highly doubt that any agent works properly at this point ... my money is still on the lobster getting better

Never goina shut up about how I'm the 803 or 804th investor in replit

@cdossman 😂😂

@dzhng Where do I send my LP checks

@GJarrosson Yea I figured 😅

@dzhng Short answer: probably not 🙃

@nearlydaniel Or you can also use duet.so - we host it as well and we also include a gateway to a bunch of common services like image / video / audio gen & enrichment & web scraping & integrations.. etc. Basically something my mom can use. No need to deal with API keys at all

I've had a lot of fun playing with OpenClaw, but there's a lot of config and tinkering needed tbh would love to set my family up but the overhead is too much — and I've been hoping someone would polish and productionize the setup. have been playing around with myclaw .ai ...

@tereza_tizkova What a run!

I resigned from E2B this week. I have been the first hire and a founding member. I’m closing a unique chapter that meant so much to me… wishing the team the best of luck!