Slow Cycling, Coffee, VC. Working on something new.
Dallas-Ft.Worth
Joined December 2008
- Tweets443
- Following1.1K
- Followers1.5K
- Likes9.7K
Raja Doddala 🇺🇸 retweeted
State of personal agents: Instinct & @bot is like giving everyone self driving vehicles but no traffic lights.
Raja Doddala 🇺🇸 retweeted
I've now had 48 hours with Instinct.
My takeway -> the moment everyone predicted just arrived.
I ran three hotel bookings, four restaurant reservations, a handful of excursions, and a number of product purchases through Instinct, all through WhatsApp.
I left voice notes. It executed.
Crazy to think not one new piece of technology was introduced. The genius: it was packaged perfectly into the right interface.
I've spent 25 years watching commerce assume a human has to arrive somewhere first. I built Magento on that assumption, from 2007 until eBay acquired it in 2011, when it ran roughly 250,000 merchants and $100B+ in annual GMV. Now I believe that assumption has stopped being true.
We've had two years of ChatGPT shopping, Perplexity commerce, agent-to-agent protocols. It went nowhere. I stayed out of investing in the space because I couldn't see the interface that would make it real.
It wasn't a model problem. It was a distribution problem, and it was solved the moment the agent appeared of inside WhatsApp.
Yes, it's a subtle nuance - but incredibly impactful.
The security and privacy questions are real. Handing an agent your card and your address raises exactly the concerns you'd expect. They'll get addressed, the same way every prior wave addressed them after adoption forced the issue, not before.
Every serious model provider will ship this within months. First mover advantage never mattered here.
The pieces were all there. We were just waiting for someone to plug them in.
Wonderful work @noahrshinn
Raja Doddala 🇺🇸 retweeted
The US has added over 1 million new jobs due to AI since 2023:
310k Blue collar:
85k Electrical contractors
85k Commercial construction
75k HVAC & plumbing
45k Utility-system construction
20k Electrical-equipment manufacturing
700k white collar:
200k Software developers
200k Maths & data science
100n Info-security analysts
80k Engineers
120k Other computer occupations
Compared to 200k losses, mostly in repetitive work like data entry and customer service.
The Economist says AI has added 1 million new jobs in America since mid-2023 (offsetting 200k AI related job losses):
“Roughly 1% of professional jobs are now ‘AI jobs’—on the order of 1m positions in America. In computer occupations and life sciences—including researchers using AI to discover new drugs—the share is 4-5%. LinkedIn’s own analysis points to roughly 640,000 new AI-specific jobs between 2023 and 2025. ‘To date, the evidence suggests that AI has been a net job creator,’ says Kory Kantenga, head of economics for the Americas at LinkedIn.”
“America’s AI infrastructure splurge has created many of them. The spending on the kit needed to make AI run—from chips and servers to data centres, cooling systems and power—is roughly $500bn a year above what it was in 2022, when the world got to know ChatGPT, calculates Goldman Sachs, a bank. Data-centre construction alone is proceeding at an annual rate of more than $75bn, nearly 60% higher than a year ago, according to Census Bureau data. That building spree requires armies of workers: electricians to wire them, HVAC specialists to stop racks from overheating, grid engineers to hook them up to the power supply and technicians to install and maintain the machines.”
“Between 2023 and 2025 employment among paralegals rose by about 11% and among market-research analysts by 6%, compared to a national average of around 2.5% (see chart 3). Despite dire warnings of imminent lay-offs, professional services are projected to keep growing rapidly thanks to demand for AI systems and consulting, according to BLS forecasts.”
Raja Doddala 🇺🇸 retweeted
- Software is *incredibly* cheap
- You have better things to do
- Like survive as a business. You are in a knife fight for your life
- What are you doing building software?
- There’s a reason there is a “software supply chain”
Most companies will realize the above, eventually
Raja Doddala 🇺🇸 retweeted
Everytime I scroll and refresh my feed it’s another post telling everyone what "real cricket" is, what formats they should watch, what they should care about, and who does or doesn't count as a "proper fan."
Stop telling people what "real cricket" is. Stop deciding which formats people should enjoy. And stop acting like someone isn't a "proper fan" because they happen to enjoy T20s, The Hundred, Tests, ODIs, franchise cricket, or any combination of the above.
Your personal preference isn't the rulebook for the rest of the sport.
The quickest way to push people away from cricket isn't a new format. It's fans constantly trying to police how other people enjoy the game.
You don't have to like every format. But you also don't get to decide what makes someone else's fandom valid.
Enjoy what you enjoy of course but stop trying to dictate everyone else's version of being a cricket fan.
Get over yourselves
Raja Doddala 🇺🇸 retweeted
I greatly admire Jayashree Ullal for her fortitude and visionary leadership. 👏👏👏
Meet Jayshree Ullal, the highest ranked Indian origin executive on Forbes' rich list, ahead of names like Sundar Pichai and Satya Nadella.
Born in London and raised in New Delhi, she trained as an electrical engineer at San Francisco State before spending over 15 years at Cisco, eventually running its data center and switching business worth $10 billion.
In 2008, she walked away from that entire empire to become CEO of Arista Networks, a startup with fewer than 50 people and $0 in revenue.
She bet on cloud networking before most people even understood what a data center was. Arista went public in 2014, and today it does close to $9 billion in annual revenue building the high speed switches that power AI infrastructure for hyperscalers like Microsoft and Meta.
Her stake in that "risky" startup is now worth an estimated $5.7 to $6.8 billion.
Everyone chasing the apps forgot somebody still has to build the plumbing underneath.
Raja Doddala 🇺🇸 retweeted
Some back of the envelope math as to why I think reports that Z.AI is running a 1GW data center using only Chinese chips are impossible:
1. China doesn't make enough chips. A 1 GW facility running Huawei's CloudMatrix systems needs ~500K Ascend 910Cs. Estimates for Huawei's domestic 910C production vary significantly, but in September 2025 @business reported Huawei made 300K Ascend 910Cs in 2025 and planned to make 600K in 2026. Using those estimates, Z.AI would have to have acquired literally every 910C China has ever manufactured domestically, and put them all in this one facility. Unlikely.
(It's also important to note that we don't know for sure China has ever made ANY Ascend 910Cs; nobody has ever torn down a single 910C that was made at SMIC, we know TSMC made millions of 910Cs for Huawei in 2023-2024, and Huawei's best AI chip that SMIC presumably is making now is worse than the 910C).
3. A 1GW data center would be the biggest data center in the world. The largest functioning data center in the world by power draw is xAI's Colossus II, which @EpochAIResearch estimates uses 0.946GW, or 946MW. Until now, the largest reported data centers in China containing only Chinese-made chips had ~10,000 chips, which would require ~18MW. China did not 50x its largest data center overnight to leapfrog the rest of the world. That's impossible.
3. Z.AI doesn't have enough money. A 1GW data center probably costs about $20 billion to stand up in China. On March 31, Z.AI reported $763 million in cash on hand, and has a $52 billion market cap. There is no way that Z.AI stood up a $20 billion data center for itself.
More likely is that China has a facility that it intends to eventually expand to 1GW, which is partially operational now. The article describes the data center as "partially operational" so this makes sense. But we have no idea how much is actually online now, or when China intends to hit that target. And this capacity is likely being rented out to many clients, including Z.AI.
But the idea that China has an operational 1GW data center is crazy, the idea that it is running on all China-made chips is even crazier, and the idea that it is all being run by one company is farcical.
Holy: China now has a 1-gigawatt AI data center built without a single Nvidia chip!
Via Bloomberg:
zAI has reportedly completed construction of the facility and begun partial operations. It runs exclusively on Chinese-made chips and will support the development of the company’s frontier GLM models.
zAI has also built or operates several computing clusters containing more than 10,000 chips each.
This is far more significant than the next Frontier model. The importance of chip independence from the USA can hardly be overstated.
