Gaurav Aggarwal

9.5K posts

Gaurav Aggarwal

Gaurav Aggarwal

@fooobar

Chief AI Scientist@Reliance Jio, Volunteer@iSPIRT. Visiting Faculty @ ISB, 2x founder, Prev: Google Research, Fashiate, Yahoo Labs. Views personal!

Bengaluru, India เข้าร่วม Eylül 2008
1.2K กำลังติดตาม16.1K ผู้ติดตาม
ทวีตที่ปักหมุด
Gaurav Aggarwal
Gaurav Aggarwal@fooobar·
A real hero from Kargil war. Seeing this video after a long long time and it still feels like his face is so familiar. Don't know what stuff folks like him are made of. Hope we are worthy of his sacrifice.
Shiv Aroor@ShivAroor

Still remember watching this interview in mesmerised silence. Capt. #VikramBatra, Param Vir Chakra. Right before the recapture of Pt.4875 on this day 21 years ago. India would change forever. He was only 24. 🇮🇳🔥

English
6
22
251
0
Gaurav Aggarwal
Gaurav Aggarwal@fooobar·
@krh_harsh Not imho, it just helps improve the product and hence gives a chance to startups (even from India) to come in!
English
0
0
0
10
Kumar Harsh
Kumar Harsh@krh_harsh·
@fooobar Does such posts/incidences cast any shadow on AI-generated code in near future?
English
1
0
0
10
Gaurav Aggarwal
Gaurav Aggarwal@fooobar·
@parrysingh Yeah, elder brother was worried if I will get a well paying job, when I decided to go for a phd :-)
English
0
0
7
445
Parminder Singh
Parminder Singh@parrysingh·
@fooobar Wow! Must have been such a radical decision at that time. Great run.
English
1
0
9
1.1K
Gaurav Aggarwal
Gaurav Aggarwal@fooobar·
18 years ago - finished my Phd in AI when even the very term was kind of unheard of outside the research circles!
Gaurav Aggarwal tweet media
English
20
11
526
28.1K
Gaurav Aggarwal
Gaurav Aggarwal@fooobar·
@bhatnaturally @ixigo Some larger technical issue probably. I also couldn't do web check-in for my air India flight this morning (from the air india app)
English
0
0
0
1.9K
bhatnaturally  🇮🇳
bhatnaturally  🇮🇳@bhatnaturally·
Let down by @ixigo’s auto web check in for the first time. Two hours to the flight don’t get the boarding pass. And then this.
bhatnaturally  🇮🇳 tweet media
English
2
0
1
2.7K
Gaurav Aggarwal
Gaurav Aggarwal@fooobar·
This was pure evil coming from a company that had "do not evil" as part of their original manifesto! They must be made to pay for this sin!
Nithin Kamath@Nithin0dha

In a landmark judgment on May 22, 2026, the Delhi High Court held Google liable for trademark infringement. The case was between Hindware and Google. The court held that, by allowing competitors of Hindware to purchase the keyword “Hindware” (a trademarked name) through Google Ads, Google enabled trademark infringement. The court said that “Hindware” is not a generic English word but a specific brand trademark. By allowing competitors to place ads on that keyword, Google is enabling competitors to divert traffic that should have legitimately gone to Hindware. This has been a big challenge for companies, both big and small. Even today, if you search for Zerodha, you will see search results from competitors. This has been happening for well over a decade. Although it is hard to quantify, we have lost a lot of business to this. Think about what happens. Whenever someone searches for "Zerodha", the traffic should rightfully come to Zerodha. But what often happens is that the first couple of results on Google Search are ads, leading the customer to a competitor's website. In the process, we lose business that should have come to us. This is made worse by the fact that we do not advertise. There is also an even more ironic thing here. A lot of brands, just to capture the traffic that should have come to them organically, end up bidding on their own keywords. Think about it. If you own a business and have a trademarked name for your business, you still have to pay Google just to hopefully make your name too expensive for your competition to run ads on it. But now, thanks to the Delhi High Court judgment, we have the option of taking legal action whenever we come across instances of other companies squatting on our keyword. The other brilliant part about this judgment is that it levels the playing field. And this matters even more for startups, who are already starved for resources and have the odds stacked against them. The last thing they need is for competitors to bid on their brand keywords and steal their traffic. This judgment now opens up a route for legal recourse whenever such deceptive practices occur. While keyword squatting is most visible in Google web results, it is an even bigger problem when it comes to app stores. Whenever someone searches for your brand, the first couple of results, both above and below your app listing, often tend to be those of your competitors. And in the case of app stores, I think the ads are even more problematic. When a user clicks on an app-store ad, they often end up installing an app. That is a much higher-commitment action than clicking on a competitor’s web search result and then just closing the page. Because the user has installed an application, the conversions, at least anecdotally, tend to be much higher. Again, brands that do not advertise are at the receiving end of this. So I welcome this ruling and hope this changes the unfair norms we've been living by for so long.

English
2
0
36
7.1K
Gaurav Aggarwal
Gaurav Aggarwal@fooobar·
So well said - "AI valuations are cross-sectionally very inefficient"
Michael Sikand 🦑@michaelsikand

Legendary stock picker @GavinSBaker ($4.1B AUM) dropped some insane bars on All In. "If you look at the valuations for all these AI [plays], they just can't all be accurate." "If the multiples on the power, cooling, optical names are correct, Nvidia and memory are going up a lot. If the multiples on Nvidia and memory are correct, everything else is probably going to underperform." "The AI market is cross-sectionally inefficient right now" Of his 10 largest positions, 8 are AI stocks. - $ALAB (8.99%) - $CIEN (6.4%) - $MU (6.26%) - $NVDA (5.31%) - $AMZN (4.84%) - $LITE (4.58%) - $GOOG (3.89%) - $COHR (3.66%)

English
0
0
2
1.4K
Gaurav Aggarwal
Gaurav Aggarwal@fooobar·
Will be in Delhi (IHC area) tomorrow (29th May) to speak at this conference. May have sometime in the evening to catch up with folks there. Feel free to ping me!
Gaurav Aggarwal tweet media
English
0
0
8
525
Gaurav Aggarwal
Gaurav Aggarwal@fooobar·
Exactly 22 minutes on the dot for today’s 5K! 4:23/km pace feels good, but that sub-22 5K is calling my name next time.
Gaurav Aggarwal tweet media
English
1
0
6
707
Gaurav Aggarwal
Gaurav Aggarwal@fooobar·
Was showing outputs of our latest TTS models to my son and he was like it sounds so natural because I can feel the breathing. Humans obviously breathe while speaking - an intangible that is probably a key ingredient that makes it sound human!
English
1
1
24
1.1K
Gaurav Aggarwal
Gaurav Aggarwal@fooobar·
To win, we need a paradigm shift. The future belongs to entirely new architectures that mimic human cognitive efficiency.
English
5
0
21
2.2K
Gaurav Aggarwal
Gaurav Aggarwal@fooobar·
Does this mean AI-driven development is a dead end? Absolutely not. It just means the current brute-force approach of throwing massive, dense frontier models at entire codebases hits a hard economic ceiling.
English
1
1
19
2.3K
Gaurav Aggarwal
Gaurav Aggarwal@fooobar·
We will enjoy cheap AI coding assistants while they last. Once VCs/tech giants stop subsidising the compute, the scaled cost of AI-generated enterprise software will likely outpace human developers. Where does the industry go when the subsidies dry up? A breakdown
Hedgie@HedgieMarkets

🦔Microsoft canceled its internal Claude Code licenses this week after token-based billing made the cost untenable, even for a company with effectively infinite cloud resources. Uber's CTO sent an internal memo warning the company burned through its entire 2026 AI budget in just four months. American AI software prices have jumped 20% to 37%, and GitHub (owned by Microsoft) is dropping flat-rate plans for usage-based billing across its products. My Take The AI subsidy era is ending in real time. The same company that put $13 billion into OpenAI and built the Azure infrastructure powering most of Anthropic's compute just looked at the bill from a competitor's coding tool and decided it was not worth paying. That is not a productivity failure on Anthropic's end. Token-based pricing is forcing every enterprise customer to confront the actual cost of running these models at scale, and the number turns out to be far higher than the flat-rate experiments suggested. This ties directly to my Gemini Flash post yesterday. Anthropic, OpenAI, and Google all raised effective prices in the last six months. Enterprises that built workflows assuming AI costs would keep falling are now watching annual budgets evaporate in months. Two outcomes look likely from here. Either enterprises scale back AI usage to fit budgets, which slows the revenue ramp the labs need to justify their valuations ahead of IPOs, or the labs cut prices and absorb the losses, which makes the unit economics worse at exactly the wrong moment. Both paths land in the same place, the numbers stop working, and somebody has to take the writedown. Hedgie🤗

English
45
34
325
86.5K