331–340 of 851 posts
Earlier quoted context omitted.
Time for my next round of Evals then. I had a 40 PR coding streak last weekend with mostly o3-mini-pro, will test the latest 2.5 now.
PR = pull request? So every bit of garbage from the LLM, over and over, resulted in an individual pull request? Why not just do one when your branch is finally right?
Earlier quoted context omitted.
Google offering Gemini 2.5 Pro for free, enough to ditch OpenAI, reminds me of an old tactic. Microsoft gained control in the '90s by bundling Internet Explorer with Windows for free, undercutting Netscape’s browser. This leveraged Windows’ dominance to make Explorer the default choice, sidelining competitors and capturing the browser market. By 1998, Netscape’s share plummeted, and Microsoft controlled access to the…
The joke's on them, because I don't have any loyalty to an LLM provider. There's very close to zero switching costs, both on the consumer front and the API front; no real distinguishing features and no network effects; just whoever has the best model at this point in time.
This article doesn't mention TPUs anywhere. I don't think it's obvious for people outside of google's ecosystem just how extraordinarily good the JAX + TPU ecosystem is. Google several structural advantages over other major players, but the largest one is that they roll their own compute solution which is actually very mature and competitive. TPUs are extremely good at both training and inference[1] especially at sca…
TPUs aren't necessarily a pro. They go back 15 years and don't seem to have yielded any kind of durable advantage. Developing them is expensive but their architecture was often over-fit to yesterday's algorithms which is why they've been through so many redesigns. Their competitors have routinely moved much faster using CUDA. Once the space settles down, the balance might tip towards specialized accelerators but NVID…
Earlier quoted context omitted.
The joke's on them, because I don't have any loyalty to an LLM provider. There's very close to zero switching costs, both on the consumer front and the API front; no real distinguishing features and no network effects; just whoever has the best model at this point in time.
The strategy worked, Netscape is no more. Eventually Google did the same to Microsoft though. I wonder if any lessons can be taken from the browser wars to how things will play out with AI models.
Earlier quoted context omitted.
The demand is there. People are already becoming addicted to this stuff.
I think the HN crowd widely overestimates how many people are even passingly familiar with the LLM landscape much less use any of the tools regularly.
Earlier quoted context omitted.
The demand is there. People are already becoming addicted to this stuff.
I think the HN crowd widely overestimates how many people are even passingly familiar with the LLM landscape much less use any of the tools regularly.
At this point in college, LLMs are everywhere. It's completely dominating history/english/mass comm fields with respect to writing papers.
Anecdotally all of my working non-tech friends use chatgpt daily.
Earlier quoted context omitted.
> F500/government are conservative and tend to stick with the vendors they know, which is why Azure has gained so much traction Outcome is the same, but being "conservative" isn't the real reason. Adding a vendor requires compliance work, process, finance etc that it's just effort. 99% of medium-large companies use Microsoft in some form so Azure can skip all of that to some extent.
That is what he meant with conservative, ie trying to not do new things because it takes more work to change.
That's not what the word conservative means, not by the dictionary or even politically.
Conservative is the averse to change or to hold traditional values without logic. It's more like a type of fear. Even if the change was easy or have 0 cost, a conservative entity won't do it.
I feel the article presents the data selectively in some places. Two examples: * The article compares Gemini 2.5 Pro Experimental to DeepSeek-R1 in accuracy benchmarks. Then, when the comparison becomes about cost, it compares Gemini 2.0 Flash to DeepSeek-R1. * In throughput numbers, DeepSeek-R1 is quoted at 24 tok/s. There are half a dozen providers, who give you easily 100+ tok/s and at scale. There's no doubt that…
This article doesn't mention TPUs anywhere. I don't think it's obvious for people outside of google's ecosystem just how extraordinarily good the JAX + TPU ecosystem is. Google several structural advantages over other major players, but the largest one is that they roll their own compute solution which is actually very mature and competitive. TPUs are extremely good at both training and inference[1] especially at sca…
But I am not sure how AWS and Google Cloud match up in terms of making this verticial integration work for their competitive advantage.
Any insight there - would be curious to read up on.
I guess Microsoft for that matter also has been investing -- we heard about the latest quantum breakthrough that was reported as creating a fundamenatally new physical state of matter. Not sure if they also have some traction with GPUs and others with more immediate applications.
Earlier quoted context omitted.
> Google is too risk averse to do anything, but completely doomed if they don't cannibalize their cash cow product. Google's cash-cow product is relevant ads. You can display relevant ads in LLM output or natural language web-search. As long as people are interacting with a Google property, I really don't think it matters what that product is, as long as there are ad views. Also: > Web search is no longer a crown jew…
Kodak was well placed to profit from the rise of digital imaging - in the late 1970s and early 1980s Kodak labs pioneered colour image sensors, and was producing some of the highest resolution CCDs out there. Bryce Bayer worked for Kodak when he invented and patented the Bayer pattern filter used in essentially every colour image sensor to this day. But the problem was: Kodak had a big film business - with a lot of f…