Tejas Srinivasan
@_Tejas_S_
Followers
963
Following
17K
Media
105
Statuses
1K
PhD student at @USCViterbi @CSatUSC. Previously research intern @allen_ai. Interested in language grounding and all things multimodal.
Joined May 2010
People are relying on AI assistance to make all kinds of decisions. *How* they incorporate AI recommendations is influenced by previous user-AI interactions and their evolving trust in the AI, which AI assistants are typically blind to. But what if they weren’t? We show that
2
18
64
We were promised flying cars and all we got was degenerate insider trading
A Google insider has officially been exposed on Polymarket. This dude just profited $1,000,000 in a single day betting on the Google search markets. Google accidentally pushed the results early, then removed them, but not before it revealed he went 22/23 on his bets and
95
1K
12K
I cannot think of a single way in which this makes society better. You're now going to have people financially - on top of emotionally - vested in their opinions which can only create further polarization. Societies need consensus, hard to see how you help achieve this when
The co-founder of Kalshi says: " The long-term vision is to financialize everything and create a tradable asset out of any difference in opinion."
229
1K
6K
Kalshi is a prediction market. Their vision is to literally apply sports betting to everything, remaking the world and our lives into a casino of speculation. This is a fundamentally dehumanizing and degenerate view of society. It must be stopped at all costs.
Tarek Mansour, CEO of Kalshi, recently stated that “the long-term vision is to financialize everything and create a tradable asset out of any difference in opinion.” People involved in such things should be dispossessed and ostracised, not celebrated as progressive icons.
142
3K
19K
to be clear, "financialize everything" means they want to find ways to generate revenue without creating anything tangible or useful. they want to profit without producing anything. that means you will, somehow, end up in debt.
The co-founder of Kalshi says: " The long-term vision is to financialize everything and create a tradable asset out of any difference in opinion."
212
5K
31K
Mamdani: Yesterday, Andrew Cuomo laughed and agreed when a radio host said that I would cheer another 9/11. Yesterday, Eric Adams said that we “can’t let our city become Europe.” He compared me to violent extremists, and he lied when he said that our movement seeks to burn
1K
7K
39K
does this remind you. of anything.
During a military-style raid on a building in Chicago’s South Shore, one resident heard a knock on his door. It wasn’t the feds — it was a mom and her 7-year-old daughter, pleading for help. He let them hide in his unit for the next 3 days. https://t.co/c9C7Zm4wsn
349
14K
73K
This is going to get me piled on a bit but a video referencing the moon landing and the fall of the Berlin Wall in a post about raising capital for a prediction market encapsulates everything I hate about the tech industry
Kalshi recently raised $300M+ at $5B from Sequoia, a16z, Paradigm and others. Since then, we've grown over 3x, hit $50B of annualized volume, and became the largest prediction market in the world. And today…Kalshi goes global. 140+ countries. 1 liquidity pool.
83
180
4K
Can someone for the life of me explain to me why the whole LLM RL community is researching GSM8K and making general claims from this? It's f*king arithmetic. I kinda got it during NeurIPS, rushed, not much better training data around. But 9 months after R1? Inexcusable.
12
3
141
if you're hanging around at the COLM workshops and are interested improving human-LLM collaboration and building reliable+accountable LLM systems, drop by the ORIGen workshop in room 518A tomorrow! https://t.co/ID44IywsDA
0
1
2
COLM attendees listening to Nicholas Carlini read excerpts from the nyt "chatgpt teen s**cide" article:
0
0
4
(Ironically, this tweet sounds like one of the sycophantic LLM responses he brought up. Fewer exclamation marks though.)
0
0
8
Nicholas Carlini with one of the best keynotes I've seen at COLM today. Not merely a survey of AI risks, but forcing LLM researchers to think with nuance about the harms we could cause to society via the tools we're building (and confront those they have already caused)
6
10
222
Gross
Largest tech campaign in LA’s history 307 billboards 160 buses 500 bus shelters $500,000
0
1
6
I'll be at #COLM2025 from tomorrow!! Hmu if you want to chat about how we can make LLMs reliable collaborators, debate whether using LLMs is rotting our brains, or want to do karaoke 🎤🎶 I'm also organizing the ORIGen workshop on Friday, come through!
0
0
3
Ever wondered what your fellow researchers think of a paper you're reading? What if you could surf through peer insights without drowning in social media chaos? Introducing SURF🏄, an interface to bring social media discussions into the paper-reading experience! #UIST2025🧵(1/n)
2
5
8
What if survey-derived rubrics 📋 graded ChatGPT instead of vibes? We benchmark LLMs & deep research systems across 75 research fields 🩺🧬🦾⚗️🏛️🎭💹: Perplexity deep research wins > 82% of head-to-heads vs the next best! w/ @realliyifei, @cmalaviya11, and @yatskar
How well can LLMs & deep research systems synthesize long-form answers to *thousands of research queries across diverse domains*? Excited to announce 🎓📖 ResearchQA: a large-scale benchmark to evaluate long-form scholarly question answering at scale across 75 fields, using
0
10
15
This is how Musk and oligarchs work: Get billions in subsidies from the federal government. Pay little or nothing in taxes. Complain about waste and fraud. Then cut federal programs that working families need. This is why we need a political revolution in this country.
SpaceX, Elon Musk’s rocket and satellite internet company, relies on federal contracts, but years of losses have most likely let it avoid paying federal income taxes, according to internal company documents.
2K
3K
11K
If you're using AI to write essays, eulogies, a text to your wife, I do think less of you as a person. Ceding your mental and creative abilities to a machine is an embarrassing thing and people should be ashamed to admit doing it in public.
380
26K
150K
Our poster is up! 🎉 We study how factuality evaluation metrics fare in their down-stream use cases uncovering "gotchas" for several NLP subcommunities. I won't be attending in person but @robinomial has graciously offered to man the post. Longer thread with details coming soon
Automatic metrics for assessing factuality are easy to run and commonly used, but do they work? In < 1 hour, come find the answer at poster 349 in Hall X4, where I’ll be presenting @ameya_godbole1 ‘s work uncovering inconsistencies, errors, and biases of factuality metrics!
0
1
19
I won't be at @aclmeeting😢 but @_Tejas_S_ will present this work on Tuesday, July 29 at 4PM in Hall 4/5! Drop by to learn about how we showed that VLMs don't actually understand human actions (<10% acc vs human's 95% acc 🤯) Here's the poster in case you miss it:
Can you tell what actions are being mimed in this video? If so, you’re smarter than AI models! Check the last tweet in this thread for answers. In a new paper, we present MIME, which evaluates whether vision language models (VLMs) have a robust understanding of human actions. 🧵
1
4
16