Saaketh Chennaiahgari Profile
Saaketh Chennaiahgari

@CSaaketh

Followers
33
Following
108
Media
5
Statuses
28

@carnegiemellon alum | fellow at @hydeparkvp | building in ai

Texas, USA
Joined July 2020
Don't wanna be here? Send us removal request.
@CSaaketh
Saaketh Chennaiahgari
1 day
Crazy to see which names ended up surviving. Also interesting to see how the companies and their logos have changed.
Tweet media one
0
0
1
@CSaaketh
Saaketh Chennaiahgari
1 day
RT @shagun_mm: 🧵1/ When people think of @CarnegieMellon, they usually think of AI and robotics. But little do they know it is also home to….
0
3
0
@CSaaketh
Saaketh Chennaiahgari
1 day
This was Silicon Valley in the ’80s. Hundreds of logos. Most forgotten. A few became legends. AI feels the same right now. The next map is being drawn. Whose name will still be on it in 2045?
Tweet media one
0
0
2
@CSaaketh
Saaketh Chennaiahgari
8 days
Everything moving at light speed with ai.
@FilmUpdates
Film Updates
8 days
Luca Guadagnino’s ‘ARTIFICIAL’ starring Andrew Garfield, Monica Barbaro and Yura Borisov reportedly begins filming this month. A recounting of the tumultuous period at artificial intelligence company OpenAI in 2023 that saw CEO Sam Altman fired and rehired in a matter of days.
Tweet media one
0
0
2
@CSaaketh
Saaketh Chennaiahgari
8 days
Big Tech used to just own the software. Now they’re owning the stack. • 2021 → $152B.• 2025 → $365B. That’s not just cloud spend. It’s physical infrastructure. AI isn’t just an application layer anymore. It’s becoming the platform—and they’re building the rails. @coatuemgmt
Tweet media one
0
0
2
@CSaaketh
Saaketh Chennaiahgari
8 days
Crazy that despite how accessible information is in this day and age, and how perfect markets seem, it’s possible for something like this to just slip under the radar for so long. For context ₹36,500 crores is almost $4 Billion.
@kirubaakaran
Kirubakaran Rajendran
9 days
What if I told you that Jane Street made ₹36,500 crores from Indian markets in just 2 years, and ₹4,800 crores of that was allegedly through market manipulation? They turned India's stock market into their personal ATM using a strategy so clever. Here's the complete details 🧵.
0
0
1
@CSaaketh
Saaketh Chennaiahgari
8 days
Too many teams are still treating AI as a feature. But 80% of AI-native companies are already deploying agentic workflows. High-growth teams? 37% of their engineers are focused on AI. If you’re not adapting, you’re already behind. (From @ICONIQCapital 2025 AI Builder’s Playbook)
Tweet media one
0
0
3
@CSaaketh
Saaketh Chennaiahgari
10 days
YC’s call for W25 is live.What’s interesting is how programs like @iCladAI are aligning with the kinds of problems YC increasingly talks about:.– Intelligent systems infra.– Safe deployment.– Long-term AI leverage.Curious to see this evolving .@sama @harjtaggar @darius @garrytan.
0
0
1
@CSaaketh
Saaketh Chennaiahgari
10 days
🔗
0
0
0
@CSaaketh
Saaketh Chennaiahgari
10 days
Donald Knuth pays $2.56 to anyone who finds an error in his work. Not for the money—for the culture. A reward for making things more correct. At YC’s AI Startup School, speakers like @karpathy, @sama @fchollet reminded us:.In AI, correctness compounds. We need more Knuth energy.
1
0
1
@CSaaketh
Saaketh Chennaiahgari
11 days
This is what happened when the Internet started unlocking value. Now look at AI. Same curve. Faster. Backed by real GDP data — and insights from YC AI Startup School (thanks to @sama, @karpathy, @AndrewYNg, @fchollet, @harjtaggar). If you missed the Internet, don’t miss AI. 👇
Tweet media one
0
1
13
@CSaaketh
Saaketh Chennaiahgari
11 days
TLDR:.• Agents are smarter workflows.• Instructions matter more than code.• Start small, test, improve.• This applies to almost any job that involves tools + decisions. Full guide here →
0
0
0
@CSaaketh
Saaketh Chennaiahgari
11 days
7. You don’t need to know how to code. You just need to understand your own workflows well enough to explain them. If you can describe how a task gets done, someone can build an agent to do it.
1
0
1
@CSaaketh
Saaketh Chennaiahgari
11 days
6. Hallucination is not failure. It’s feedback. The guide treats bad outputs as debugging clues—not dead ends. Small changes in tool access, instruction wording, or memory often fix it.
1
0
0
@CSaaketh
Saaketh Chennaiahgari
11 days
5. Good agents think in steps. One surprising insight: agents that plan their next step before acting are far more accurate. It’s not about speed—it’s about control.
1
0
0
@CSaaketh
Saaketh Chennaiahgari
11 days
4. The future isn’t multi-agent everything. OpenAI warns against overengineering. Most great agents today are simple loops:.“Try something → See what happened → Try again.”.One smart agent > ten confused ones.
1
0
0
@CSaaketh
Saaketh Chennaiahgari
11 days
3. You already use agents—you just don’t call them that. Examples from the guide:.• Refund processors.• Intake assistants.• Knowledge search bots.If it handles steps, tools, and adapts—it’s an agent.
1
0
0
@CSaaketh
Saaketh Chennaiahgari
11 days
2. The hardest part isn’t the AI. It’s the instructions. OpenAI shows that well-written instructions (like job training docs) often matter more than the model itself. The agent is only as good as what you teach it.
1
0
0
@CSaaketh
Saaketh Chennaiahgari
11 days
1.Agents don’t replace people—they replace processes. Think less “robot employee,” more “automated teammate” that knows how to get from A to B using judgment, not just scripts.
1
0
0
@CSaaketh
Saaketh Chennaiahgari
11 days
OpenAI just released a guide on how to build AI agents. But it’s not just for engineers. If you run a business, work with customers, or do anything repetitive—this matters. Here are the deeper lessons most people will miss 👇.
1
0
2