Karthik (@krthkk) header
Karthik (@krthkk) profile picture

build and create | AI journal with voice transcription to extract insights - ThoughtsAI: https://teapawt.net/t.co/9MOAIN6DCY

Malaysia
Joined March 2009
Photo by Karthik (@krthkk)Photo by Karthik (@krthkk)Photo by Karthik (@krthkk)Photo by Karthik (@krthkk)Photo by Karthik (@krthkk)Photo by Karthik (@krthkk)Photo by Karthik (@krthkk)Photo by Karthik (@krthkk)Photo by Karthik (@krthkk)Photo by Karthik (@krthkk)
A ‘hello world’ milestone for Project Suncatcher 🚀 After launching our prototype satellite a few weeks ago, we’ve now successfully sent a command up to orbit, booted up the TPU, ran the prompt, and received our first complete response back on earth - a haiku!
1,477
2,710
1,279
26,669
4,376,888
the physics behind the roadster - roadster.pplx.app/
103
362
46
2,893
239,158
Karthik retweeted
People on X keep laughing at Google's AI strategy. I actually think they are playing their cards really well. Google Search was the first big tech product under real threat from AI. And it is their main business. Instead of burning compute on oversized benchmark models like everyone else, they bet heavily on Flash. Smaller, faster, and cheap enough to serve billions. Then they put it right inside their strongest existing products. That is how they saved their core business, using the same "dumb" overview model everyone keeps complaining about. While other tech giants lose money subsidizing wild compute costs, Google used that compute to protect their castle. In reality, they might be the only ones making real money with AI right now. Meanwhile, DeepMind quietly works on RSI in the background with the rest of the compute. Honestly, it is a smart play.
144
58
17
1,209
87,393
Gemini 3 Deep Think is now available for Google AI Ultra subscribers in the @GeminiApp, incorporating our gold medal winning IMO and ICPC technologies! 🏅With its parallel thinking capabilities it can tackle highly complex maths & science problems - enjoy!
178
275
58
3,276
342,099
Sam Altman says OpenAI is severely short compute. Tune out the noise. We are going higher.
73
45
8
889
85,051
There is incredible alpha in learning how things work and why. And why people chose to build systems the way they did. That’s an underrated part to making something new. The future will not be the same as the past. But it will rhyme. And knowing how it works is the first step to making the right changes that fit today’s tech and today’s market needs.
65
112
11
1,037
78,925
Karthik retweeted
Replying to @_mctrinh
RajaSudoku on iOS for my kid and myself :) apps.apple.com/my/app/rajasu… Would love you to try and let me know what you think of it!
1
10
388
Gemini 2.5 Pro is terrifyingly good at real tasks. But most people don’t know what to do with it. I used it to automate research, content, code reviews, and more. Here are 10 ways to use Gemini 2.5 Pro and automate your tedious work:
38
98
4
1,191
301,742
Demis is a genius and he should be given all resources in the world to make this happen.
85
235
44
3,240
271,684
Karthik retweeted
there’s a strong case to be made for mr milchick being one of the best characters in recent memory
347
3,460
1,174
51,007
3,656,088
We’re rolling out AI Mode to Google One AI Premium subscribers today, opt in on Labs. And just like AI Overviews, AI Mode will get better with time and feedback. Get details here: blog.google/products/search/…
31
80
12
458
83,735
AI Overviews are one of our most popular Search features ever. They keep getting better as models improve, and now, Gemini 2.0 is powering responses for coding, advanced math, and multimodal queries in the US, with more to come. Today, we’re also introducing our newest Labs experiment for Search: AI Mode. You’ll get AI responses using Gemini 2.0’s advanced reasoning, thinking, multimodal capabilities + new ways to explore even more of the web.
474
317
160
3,190
683,296
Google AI Mode vs Perplexity. Google is a giant for a reason. The answer is so accurate by Google AI Mode
1
4
516
I was given early access to Grok 3 earlier today, making me I think one of the first few who could run a quick vibe check. Thinking ✅ First, Grok 3 clearly has an around state of the art thinking model ("Think" button) and did great out of the box on my Settler's of Catan question: "Create a board game webpage showing a hex grid, just like in the game Settlers of Catan. Each hex grid is numbered from 1..N, where N is the total number of hex tiles. Make it generic, so one can change the number of "rings" using a slider. For example in Catan the radius is 3 hexes. Single html page please." Few models get this right reliably. The top OpenAI thinking models (e.g. o1-pro, at $200/month) get it too, but all of DeepSeek-R1, Gemini 2.0 Flash Thinking, and Claude do not. ❌ It did not solve my "Emoji mystery" question where I give a smiling face with an attached message hidden inside Unicode variation selectors, even when I give a strong hint on how to decode it in the form of Rust code. The most progress I've seen is from DeepSeek-R1 which once partially decoded the message. ❓ It solved a few tic tac toe boards I gave it with a pretty nice/clean chain of thought (many SOTA models often fail these!). So I upped the difficulty and asked it to generate 3 "tricky" tic tac toe boards, which it failed on (generating nonsense boards / text), but then so did o1 pro. ✅ I uploaded GPT-2 paper. I asked a bunch of simple lookup questions, all worked great. Then asked to estimate the number of training flops it took to train GPT-2, with no searching. This is tricky because the number of tokens is not spelled out so it has to be partially estimated and partially calculated, stressing all of lookup, knowledge, and math. One example is 40GB of text ~= 40B characters ~= 40B bytes (assume ASCII) ~= 10B tokens (assume ~4 bytes/tok), at ~10 epochs ~= 100B token training run, at 1.5B params and with 2+4=6 flops/param/token, this is 100e9 X 1.5e9 X 6 ~= 1e21 FLOPs. Both Grok 3 and 4o fail this task, but Grok 3 with Thinking solves it great, while o1 pro (GPT thinking model) fails. I like that the model *will* attempt to solve the Riemann hypothesis when asked to, similar to DeepSeek-R1 but unlike many other models that give up instantly (o1-pro, Claude, Gemini 2.0 Flash Thinking) and simply say that it is a great unsolved problem. I had to stop it eventually because I felt a bit bad for it, but it showed courage and who knows, maybe one day... The impression overall I got here is that this is somewhere around o1-pro capability, and ahead of DeepSeek-R1, though of course we need actual, real evaluations to look at. DeepSearch Very neat offering that seems to combine something along the lines of what OpenAI / Perplexity call "Deep Research", together with thinking. Except instead of "Deep Research" it is "Deep Search" (sigh). Can produce high quality responses to various researchy / lookupy questions you could imagine have answers in article on the internet, e.g. a few I tried, which I stole from my recent search history on Perplexity, along with how it went: - ✅ "What's up with the upcoming Apple Launch? Any rumors?" - ✅ "Why is Palantir stock surging recently?" - ✅ "White Lotus 3 where was it filmed and is it the same team as Seasons 1 and 2?" - ✅ "What toothpaste does Bryan Johnson use?" - ❌ "Singles Inferno Season 4 cast where are they now?" - ❌ "What speech to text program has Simon Willison mentioned he's using?" ❌ I did find some sharp edges here. E.g. the model doesn't seem to like to reference X as a source by default, though you can explicitly ask it to. A few times I caught it hallucinating URLs that don't exist. A few times it said factual things that I think are incorrect and it didn't provide a citation for it (it probably doesn't exist). E.g. it told me that "Kim Jeong-su is still dating Kim Min-seol" of Singles Inferno Season 4, which surely is totally off, right? And when I asked it to create a report on the major LLM labs and their amount of total funding and estimate of employee count, it listed 12 major labs but not itself (xAI). The impression I get of DeepSearch is that it's approximately around Perplexity DeepResearch offering (which is great!), but not yet at the level of OpenAI's recently released "Deep Research", which still feels more thorough and reliable (though still nowhere perfect, e.g. it, too, quite incorrectly excludes xAI as a "major LLM labs" when I tried with it...). Random LLM "gotcha"s I tried a few more fun / random LLM gotcha queries I like to try now and then. Gotchas are queries that specifically on the easy side for humans but on the hard side for LLMs, so I was curious which of them Grok 3 makes progress on. ✅ Grok 3 knows there are 3 "r" in "strawberry", but then it also told me there are only 3 "L" in LOLLAPALOOZA. Turning on Thinking solves this. ✅ Grok 3 told me 9.11 > 9.9. (common with other LLMs too), but again, turning on Thinking solves it. ✅ Few simple puzzles worked ok even without thinking, e.g. *"Sally (a girl) has 3 brothers. Each brother has 2 sisters. How many sisters does Sally have?"*. E.g. GPT4o says 2 (incorrectly). ❌ Sadly the model's sense of humor does not appear to be obviously improved. This is a common LLM issue with humor capability and general mode collapse, famously, e.g. 90% of 1,008 outputs asking ChatGPT for joke were repetitions of the same 25 jokes​. Even when prompted in more detail away from simple pun territory (e.g. give me a standup), I'm not sure that it is state of the art humor. Example generated joke: "*Why did the chicken join a band? Because it had the drumsticks and wanted to be a cluck-star!*". In quick testing, thinking did not help, possibly it made it a bit worse. ❌ Model still appears to be just a bit too overly sensitive to "complex ethical issues", e.g. generated a 1 page essay basically refusing to answer whether it might be ethically justifiable to misgender someone if it meant saving 1 million people from dying. ❌ Simon Willison's "*Generate an SVG of a pelican riding a bicycle*". It stresses the LLMs ability to lay out many elements on a 2D grid, which is very difficult because the LLMs can't "see" like people do, so it's arranging things in the dark, in text. Marking as fail because these pelicans are qutie good but, but still a bit broken (see image and comparisons). Claude's are best, but imo I suspect they specifically targeted SVG capability during training. Summary. As far as a quick vibe check over ~2 hours this morning, Grok 3 + Thinking feels somewhere around the state of the art territory of OpenAI's strongest models (o1-pro, $200/month), and slightly better than DeepSeek-R1 and Gemini 2.0 Flash Thinking. Which is quite incredible considering that the team started from scratch ~1 year ago, this timescale to state of the art territory is unprecedented. Do also keep in mind the caveats - the models are stochastic and may give slightly different answers each time, and it is very early, so we'll have to wait for a lot more evaluations over a period of the next few days/weeks. The early LM arena results look quite encouraging indeed. For now, big congrats to the xAI team, they clearly have huge velocity and momentum and I am excited to add Grok 3 to my "LLM council" and hear what it thinks going forward.
655
2,154
681
16,564
3,737,384
Replying to @ecommerceshares
Absolutely agree on this. At a minimum, they could have created a Cursor-like tool which integrates directly in VS Code and works in a PCI compliant environment (Azure has the compliance certificate). Since the tool is fully integrated, allow it to work with 100% of an existing code base, and set it up so that you can just write feature specs that it codes + tests for you. Take it a step further for newbies or greenfield projects by giving people a wizard that can work with them on building projects. Give people starter credits and Have it deploy into the Azure ecosystem. Have it scale with them too (eg for only a little more per month you can get logging, telemetry, etc.). Frictionless deployments would keep people (especially new ones who have been able to code thanks to your tooling) sticky with your software. This was a massive product miss which could have easily been solved.
17
4
2
159
44,567
I’m still amazed how Microsoft fucked up what was a clear first mover advantage in AI by simply making terrible product decisions. A failure of product, marketing, positioning… everything. Organisational disaster.
344
192
56
5,134
950,089
Replying to @aaditsh
The videos are being added here: youtube.com/@CS153InfraTeam/…
8
74
6
1,050
53,675
Someone please open-source this. Everyone in the world deserves high quality education.
57
349
38
8,687
568,183
At current pace, my kids are on track to finish high school academics by age 13. That's with two hours per day of personalized online learning for less than $30/mo.
267
659
88
14,002
1,710,253
It’s amazing that no model has caught up with Claude on coding. Even if they look good on benchmark they’re still not as good at generating working good looking modern web apps. Whatever magic Anthropic did seems very durable.
242
276
65
5,901
561,288