Ianjit

6 ms·

31.
by Ianjit 1y ago
Matt Levine never worked in crypto, but he knew SBF/FTX was a scam. Humans are humans.
32.
by Ianjit 1y ago
METR's study uses a 50% sucess rate. For most enterpise applications a 50% pass rate is unacceptable in automation, more like >99.9%.
33.
by Ianjit 1y ago
METR uses a 50% success rate in that analysis, beecause the models are non-determistic.
34.
by Ianjit 1y ago
For reasoning the data suggests models are in the logistic domain? Grok 3 - Grok 3 reasoning: 15% increase in training compute for a 25% uplift in inteligence Grok 3 reasoning - Grok 4: 80% increase in training compute for a 15% uplift in i
35.
by Ianjit 1y ago
Conor Benn?
36.
by Ianjit 1y ago
"That study design has some issues. " This is a study that tries to be scientific, unlike developer self reports and CEO promises of 10x. Can you point to a better study on the impact of AI on developer productivity? The only othe
37.
by Ianjit 1y ago
"Just start iterating in chunks alongside the LLM". For those types of tasks it probably takes the same amount of time to form the idea without AI as with AI, this is what Metr found in its study of developer productivity. https:
38.
by Ianjit 1y ago
Just recycle the headlines from 5 years go: https://www.nature.com/articles/s41586-021-03544-w.epdf
39.
by Ianjit 1y ago
"expect an LLM to correctly produce the complex thing they have in mind" My guess is that for some types of work people don't know what the complex thing they have in mind is ex ante. The idea forms and is clarified through
40.
by Ianjit 1y ago
"If you build too slowly and superintelligence is possible in 3 years, but you built it out assuming it is possible in 5 years, then you are out of position on the most important technology" This made me think Zuck sees it as a qu
41.
by Ianjit 1y ago
I do. It typically goes on to write a preamble about why it gave a long answer before finally providing a summary. Token stuffing.
42.
by Ianjit 1y ago
Zuck has made it clear this is a Pascal's Gamble: "If we end up misspending a couple of hundred billion dollars, I think that is going to be very unfortunate, obviously. But I actually think the risk is higher on the other side .
43.
by Ianjit 1y ago
I notice that when I ask ChatGPT a question the answers I get back seem more verbose than they were a year ago. Where a 1-2 line response would be sufficient GPT delivers a sprawling essay. Overall the value of the answer has probably gone
44.
by Ianjit 1y ago
Big Tech will find a way of ruining it for you.
45.
by Ianjit 1y ago
The clip length is 20 seconds.
46.
by Ianjit 1y ago
The TAM for TV/Film storyboard is probably way too small to justify the cost of training and infering these models.
47.
by Ianjit 1y ago
Free user inference is probably accounted for in Sales and Marketing. Adjusted COGS.
48.
by Ianjit 1y ago
Provide verbose answers, increases tokens. Demand is measured in tokens, so it looks like demand is sky rocketing. Valuation goes up. I have noticed that GPT now gives me really long explanations for even the simplest questions. Thankfully
49.
by Ianjit 1y ago
Are Azure/GCP making a profit with AI? ORCL definatly isn't, FCF will go heavily negative.
50.
by Ianjit 1y ago
The incremental compute costs will scale with the incremental data added, therefore training costs will grow at a much slower rate compared to when training was GPU limited.
51.
by Ianjit 1y ago
Users are waiting for Blackwell. Then Rubin. CRWV depreciates GPUs over 6 years. Rails last a lot longer.
52.
by Ianjit 1y ago
I am at stage 4 as well. Boredom.
53.
by Ianjit 1y ago
Dot com was a financial bubble, but the internet was still very usefull. Financial markets can become (often are) dislocated from reality.
54.
by Ianjit 1y ago
Parallel vs serial computation... its something you would hope software engineers understand...
55.
by Ianjit 1y ago
Or to assuage investors participating in the OpenAI secondary on the issue of cash burn.
56.
by Ianjit 1y ago
5-10 years and Radiologists will be out of a Job, just you wait and see.
57.
by Ianjit 1y ago
Historically there has been a gap between the performance of AI in test environments vs the impact in the real world, and that makes people who have been through the cycle a few times cautious extrapolating. In 2016 Geoffrey Hinton said vis
58.
by Ianjit 1y ago
You don't have to short the bubble to watch it pop.
59.
by Ianjit 1y ago
ORCL is guiding to $114bn in OCI revenue for FY29 (YE May), that is more than AWS delivered in 2024. ORCL's FY29 revenue guide for OCI is basically the same as OpenAI's revenue target (as reported by the information) for the same
60.
by Ianjit 1y ago
I'm sorry, I misunderstood your question, I didn't see where it was asking about being reasonable. "As individuals? If your friends and family all use meta products, are you suggesting to get new friends and family, or to con

More ›