madoit
@madooit
Followers
358
Following
767
Media
68
Statuses
1K
AI venture studio @icecreamlabs 10+ yrs of AI geek via Stanford + Virginia tech + lot of cities. @icecreamlabs
Joined November 2008
‘Data-driven’ = using metrics to make decisions… as long as no one asks which metric. 🤔📈
0
1
2
Making AI work with world data on real world problems is hard. #ai #MachineLearning
0
0
0
The biggest winners of the weekend: Microsoft Google meet Twitter / X All the parodies and humour on X
0
0
0
Amazing how many founders I meet have had not customer contact even. Have been there as an engineer in my first startup but learnt my lessons the hard way.
Please save your startup from a lot of pain and talk to customers before wasting time solving a problem that does not exist.
0
0
1
Apart from the impact on @sama & the folks at @OpenAI. This has huge shock for broader GenAI ecosystem and community building on openai. Creating opportunity and urgency for alternatives from @Google @AIatMeta @grok @MistralAI @antrophic
0
0
1
Huggingface is still under 300 people and Anthropic is sub 200. Still relatively small.
Interesting how some top Generative AI have very small teams * Character AI: 30-40 employees * Inflection AI: 40 employees * Midjourney: 70-80 employees (was much smaller for a while) Yes, top folks, and team sizes will grow, but that’s still a lot of impact per employee
0
0
0
I am looking for a #magento freelancer who can come help me out at my office setting up my website and train me - website to be launched soon. Development work has been done. Office at Jeevanbhima Nagar #Bangalore - needless to say, startup budget. RT for #socialkarma pl 🙏.
1
1
2
Elevate your image game with https://t.co/bppDtLgltN! Whether it’s for personal or professional use, our cutting-edge technology will make your images stand out like never before. Try it now! #Breeze #ImageEnhancement #Technology #StandOutImages
0
2
0
4. Validate your solution. Go/No go decision. Take demo to 10 users. Validate problem still exists and ur solution hits the mark. Go back to few previously interviewed. Be brutally honest to make go/no-go decision. You will be so much wiser and smarter.
0
0
0
3. Quick MVP/ demo solution: Make a list of essential features from your ideas, user needs, competitors. Prioritise to cover v1 only. Build a demo/proto/mvp. My go to figma. Easy, cheap. $1000 Will help you crystallise your ideas / features.
1
0
0
2. Goal: go/no go decision. Be brutal honest. Get market, competitor data online. You validate problem, buyer pricing, gtm options. Map user needs to ur ideas. Quick market size, any friction gtm.
1
0
0
1. Interview: Even if you know problem - talk to atleast 5 (ideally 10+) diverse potential users to validate problem different type of users/companies. Not just your buddies/colleagues/friends. Use design thinking. Ask for 30m slots.
1
0
0
So what is back log? I just let it track all emails and never update calls 😀
Anyone else tragically allergic to keeping their Hubspot up to date? I have to do these huge batches to keep myself organized...would be *so easy* if I just added leads and updated statuses in the moment.
0
0
0
Big news from Openai: Function calling-one step closer to integrating LLMs directly with 3rd party apps & systems. Soon we can instruct other apps directly from chatgpt. Amazed at their execution. https://t.co/O9pVJpfLCM
0
0
0