ashwin_gupta Profile Banner
Ashwin Gupta Profile
Ashwin Gupta

@ashwin_gupta

Followers
479
Following
957
Media
58
Statuses
914

Strategy @ Zomato | 1 of 18 hired from 18K+ via Deepinder’s CoS program | 2x Founder | Last startup: $150k+ ARR | VWO: $18M → $50M ARR | 12+ yrs in Product

New Delhi, India
Joined March 2009
Don't wanna be here? Send us removal request.
@deepigoyal
Deepinder Goyal
11 months
Announcement -
348
115
2K
@artificialguybr
Artificial Guy / João Vitor A.
2 years
Google raised the price of Gemini 1.5 Flash output by 98% without telling anyone. This just a week after announcing the model. Output goes from 0.53/1M to 1.05/1M. How can we trust an API that drastically increases its price in a week? lol
56
72
709
@ashwin_gupta
Ashwin Gupta
2 years
What are some of the best products, both open and closed source, based on agentic workflows out there? #ai #llm
0
0
2
@ashwin_gupta
Ashwin Gupta
2 years
"Sora, please generate a version of Game of Thrones Seasons 6-8 where they didn't suck." __________________ First video created by @OpenAI's Sora AI 👇🏼 #openai #ai #music #films #sora #chatgpt4o
0
0
2
@ashwin_gupta
Ashwin Gupta
2 years
Here's the most important takeaway from @nvidia's conference call last night: The next industrial revolution has begun. Companies and countries are partnering with NVIDIA to shift traditional data centers to AI factories, producing artificial intelligence. This shift is expected
0
0
0
@ashwin_gupta
Ashwin Gupta
2 years
It's not her or 'her'. It's clearly another timbre of voice. I don't get what the fuss is about. #chatgpt4o #openai #her #sky
0
0
1
@ashwin_gupta
Ashwin Gupta
2 years
Here's the working link:
drive.google.com
0
0
0
@ashwin_gupta
Ashwin Gupta
2 years
Managers/leaders think bigger teams mean faster work. WRONG. It's like saying if one Tesla can go 0 to 60 in 3 seconds, then 10 Teslas can do it in 0.3 seconds. Makes no sense, right? The research paper I've shared calls it the "team scaling fallacy," and frankly, it's a
1
0
0
@ashwin_gupta
Ashwin Gupta
2 years
Facebook Messenger's Highlights tab is the Facebook we were promised in 2006. It was pure, social magic focused on people around you, before the news feed nonsense ruined everything. The endless scroll of garbage turned an incredible platform into a mess. Sigh. #facebook
0
0
1
@ashwin_gupta
Ashwin Gupta
2 years
@IntuitMachine @bindureddy what do you guys think about this?
0
0
0
@mwseibel
Michael Seibel
2 years
An easy founder mistake is to treat your high quality employees as children. It's easier to tell people what to do and how to think but the best people will not thrive in this sort of environment. A second easy mistake is to employ low quality employees. They must be told what
36
122
860
@ashwin_gupta
Ashwin Gupta
2 years
Imagine a social network where AI agents are the first million users. These agents generate intriguing posts, engage in dynamic interactions, and create an endless stream of compelling content. Humans join in, becoming part of vibrant conversations, learning from and contributing
1
0
0
@ashwin_gupta
Ashwin Gupta
2 years
Thinking of starting a @substack that summarises 5 new AI papers in 3 points, everyday. Thinking of calling it in3beats. Thoughts? Would you be interested in something like this? #ai #buildinpublic
1
0
1
@ashwin_gupta
Ashwin Gupta
2 years
Where's the Twitter thread that explains how to build LLMs from scratch!? Asking for a friend.. #ai #llm #chatgpt #openai
0
0
0
@ashwin_gupta
Ashwin Gupta
2 years
For startups born in the AI era, prompting is the new hiring. I mean why the hell would you hire when you can code an AI to get the job done? Save time, save money, and let algorithms do the heavy lifting. #buildinpublic #AI #chatgpt #FutureOfWork
1
0
5
@ashwin_gupta
Ashwin Gupta
2 years
Can anyone help with this? #gpt4o
@kalradivyanshu
Divyanshu Kalra
2 years
Does anyone know how @OpenAI is measuring the reported 320ms average latency of GPT-4o? I measured the time between the time the human stopped speaking and GPT started speaking in the demo video, and it is always 900ms+? Why the discrepancy?
3
0
0
@ashwin_gupta
Ashwin Gupta
2 years
Terrifying.
@anitaleirfall
Anita Leirfall
2 years
Brain-Reading Device Deciphers Internal Thoughts With Surprising Precision Scientists have created a device that can read a person's brain signals and translate their internal speech, at least on a modest level https://t.co/8mToSBsD8r via @gizmodo (h/t @GreggDCaruso)
1
0
2
@ashwin_gupta
Ashwin Gupta
2 years
I tried Gemini Flash 1.5. It's super fast. However, not all the time. I gave it a system prompt of 3k tokens and suddenly it started taking >2 seconds to process answers. In comparison, llama 3 on groq is a couple of hundred milliseconds. #Googleio #llama
2
0
1
@ashwin_gupta
Ashwin Gupta
2 years
I tried making chatgpt4o read a pdf file but it couldn't and it didn't have any text. I thought it was multi-modal? Shouldn't it be able to "read" it? What am I missing? #chatgpt4o
0
0
0
@ashwin_gupta
Ashwin Gupta
2 years
I never thought I'd say this, but Google needs to just... ship. Slap that beta label on it and launch already, brother 🚀 #GoogleIO #GoogleIO2024 #OpenAl
1
0
2