Umesh Gangundi (@umesh_g) header
Umesh Gangundi (@umesh_g) profile picture

Let’s talk about LLMs and agents Routine: Explore - Code - Share - Repeat; Interests: tech, startups, movies, books, stock market, photography

Auckland, New Zealand
Joined December 2009
Umesh Gangundi retweeted
We’ve written an FAQ to answer some of the questions we've received about watermarking. In summary: • We’re implementing watermarking to comply with the EU AI Act. Other major model developers have signed the same Code of Practice and will also be implementing watermarking; • Our watermarking method doesn’t have any practical impact on the quality or content of Claude’s outputs; • The difference between watermarked and un-watermarked text will not be distinguishable to readers; • Nothing is added to the text and there are no hidden characters; • Watermarking doesn’t require extra tokens, and will not be more expensive; • Watermarks can’t be traced to a specific person, organization, or chat. Read more: anthropic.com/news/claude-te…
1,895
675
1,212
4,816
11,327,789
Umesh Gangundi retweeted
For my first post, I’m sharing a letter @NVIDIA signed on why open models matter. AI will transform every industry, power every company, and be built by every country. Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty. The world needs both frontier closed models and frontier open models. images.nvidia.com/pdf/Open-W…
16,564
34,464
12,076
193,881
69,437,704
Whenever there is new announcement or a bug fix, most of the llm providers resetting the sessions for every user. Now I’m sitting and thinking, how much actually tokens cost, #overpriced or #underpriced ? #llm
1
2
32
#HyderabadRains Rain of 4cms in this season, City is on the knees . Unplanned growth + removal of FSI and concentration of companies at just west of Hyderabad. In this same area recently one acre of Govt land was sold for 240 Crores
17
HTML export of the agent output for visual consumption is super useful. Give it a try ref: thariqs.github.io/html-effec… Thanks @trq212
9
We need more of small target domain models than a large generic foundational models #llm #sllm #AIonEdge #food #recipes
Launching our new paper on arXiv: we trained the largest multilingual food model ever built. 4.1M recipes. 7 languages. 1,790 ingredients. 300 dimensions. All of human cooking compressed into 2 megabytes.
1
1
2
84
It’s been a while on here, but I’m going to make a regular habit of posting about LLMs and AI and my experiences going forward on this account. #reset
22
Umesh Gangundi retweeted
9
1,159
58
4,589
169,954
Excited to share my latest blog post exploring #dalle3 and #midjourney image generation! 🚀 #GenerativeAI #prompts blog.umess.me/2024/01/05/dal…
1
47
Umesh Gangundi retweeted
Drunk Delhi Dude asking the Lion if he knows of his father
326
918
201
6,823
1,025,449
Umesh Gangundi retweeted
iMessage honest ETA
299
2,256
3,054
32,568
6,732,477
What could be consideration around having #UPIlite limit as 200? 🤔
50
Umesh Gangundi retweeted
New comic! Connection
3
83
3
287
23,067
Umesh Gangundi retweeted
Ready to take off with 20 Boeing 787s and 10 Boeing 777-9s widebody aircraft, and 190 Boeing 737 MAX single-aisle aircraft. The B777/787s will be powered by GE Aerospace and B737 Max by CFM International #ReadyForMore @BoeingAirplanes @GE_Aerospace @GEIndia @CFM_engines
80
483
33
4,722
224,270
Umesh Gangundi retweeted
AI is committed towards its transformation journey. As a part of the same, we are celebrating the order of 470 aircraft with @Airbus @BoeingAirplanes @RollsRoyce @GE_Aerospace @CFM_engines
287
1,465
233
9,437
758,182