keeda

42 ms·

631.
by keeda 9mo ago
You're right, maybe the better word would have been "presumably."
632.
by keeda 9mo ago
I would hope so, but it's not yet clear if the economics pan out for large scale deployment. The ride is amazing, but the sensor-laden cars are also very expensive. The only sensible aspect of Elon's boneheaded move to remove non-
633.
by keeda 9mo ago
There were blogs that explained this in detail (Facebook does something similar), but I can't find them, so here's what Google's AI overview says when I search for "How gmail cookies help google track users across the we
634.
by keeda 9mo ago
And potentially their investors, for the startups that got funding. And maybe their customers, for the startups that are already hitting millions in ARR in a year or less. There are some of the numbers out there worth looking up. Is your co
635.
by keeda 9mo ago
This will sound pedantic but I'll explain why it matters: the exact wording is "We need to do something useful with AI", because the current (editorialized) submission title (of an already slanted article) makes it sound li
636.
by keeda 9mo ago
That is also the thesis of this piece: https://www.wheresyoured.at/the-men-who-killed-google/ It is plausible, but I'd guess Google would not risk that. I'm sure Google has pulled other shenanigans to get mor
637.
by keeda 9mo ago
Google's advantage is not just in its index and algorithms, it is that it has built a self-reinforcing flywheel that data mines human attention at massive scale to improve their search results. This comment ( https://news.y
638.
by keeda 9mo ago
> I am making the point that Google effectively benefited from the large amount of human labor... This is exactly right, but the thing most people miss is that Google has been using human intelligence at massive scale even to this day
639.
by keeda 9mo ago
Flattery is a form of sycophancy ;-)
640.
by keeda 9mo ago
Technically correct, but the workarounds AI search engines use for grounding results could be a close enough approximation. Might not be accurate, but could be better than nothing. Also Anthropic is doing interesting work in interpretabilit
641.
by keeda 9mo ago
> Are they really late? Has everyone started using agents and paying $200 subscriptions? If you rephrase the question as "Are most engineers already using AI?" -- because it transcends the specific modality (agents vs chat vs
642.
by keeda 9mo ago
Yes, the world model building is achieved via pattern matching and happens during ingestion and training, but that is also part of the intelligence.
643.
by keeda 9mo ago
Yes, it could be that intelligence is essentially a sophisticated form of recursive, brute force pattern matching. I'm beginning to think the Bitter Lesson applies to organic intelligence as well, because basic pattern matching can be
644.
by keeda 9mo ago
The dynamics are exactly right, but I would say engineers can't "let" "politically bad" projects fail, because fixing that is, in a very real sense, above their paygrade. That responsibility lies with the executives
645.
by keeda 10mo ago
Or, if you focus on the "slop" aspect of AI, the bland leading the bland ;-)
646.
by keeda 10mo ago
It's not just your bias, I too have found great success with a functional programming style, even from the earliest days of ChatGPT. (Not Haskell, but JS, which the models were always good at.) I think the underlying reason is that fun
647.
by keeda 10mo ago
What "domain expert" means is also changing however. As I've mentioned often, I'm solving problems in a domain I had minimal background in before. However, that domain is computer vision. So I can literally "see&quo
648.
by keeda 10mo ago
Yeah unfortunately the hype is overwhelming and it needs real work to figure out what the real impact is. At this point the gains are modest but still compelling. On the other hand, we are still going through a period of discovering how to
649.
by keeda 10mo ago
One of the biggest shortcomings of Open Source was that it implicitly defaulted to a volunteer model and so financing the work was always left as an exercise for the reader. Hence (as TFA points out) open source code from commercial entitie
650.
by keeda 10mo ago
As the Wikipedia page calls out, predatory pricing is generally in the context of a dominant firm throwing their weight around to dominate the market. You could make this case against large incumbents like Microsoft and Google, but Anthropi
651.
by keeda 10mo ago
I mention a few here: https://news.ycombinator.com/item?id=45379452 > ... just looking at LOC or PRs, which of course is nonsense. That's basically a variation of "How can they prove anything when we don&#x27
652.
by keeda 10mo ago
Why?
653.
by keeda 10mo ago
If the software has tens of thousands of users without expecting to get any at all, does the code even matter?
654.
by keeda 10mo ago
It does, but I have no mental model of what would be required to efficiently coordinate a bunch of independently operating agents, so it's hard to make a judgement. Also about half of it seems to be tests. It even has performance bench
655.
by keeda 10mo ago
Actually, quite the opposite. It seems any positive comment about AI coding gets at least one response along the lines of "Oh yeah, show me proof" or "Where is the deluge of vibe-coded apps?" For my part, I point out the
656.
by keeda 10mo ago
Yeah, it somewhat is! It also made some mistakes analogous to what psychics would based on the limited sample of exposure it had to me. For instance, I've been struggling against a specific problem for a very long time, using ChatGPT h
657.
by keeda 10mo ago
Ah, I guess I've been interpreting "moat" narrowly, such as, keeping your competitors from muscling in on your existing business, e.g. siphoning away your existing users. Makes sense that it applies in the broader sense as we
658.
by keeda 10mo ago
The memory is definitely sort of a moat. As an example, I'm working on a relatively niche problem in computer vision (small, low-resolution images) and ChatGPT now "knows" this and tailors its responses accordingly. With othe
659.
by keeda 10mo ago
These are important conversations to have because there is so much hyperbole in both directions that a lot of people end up having strong but misguided opinions. I think it's very helpful to consider the impact of LLMs in context (hehe
660.
by keeda 10mo ago
These tools may not be turning a profit yet, but as many point out, this is simply due to deeply subsidized free usage to capture market share and discover new use cases. However, their economic potential is undeniable. Just taking the exam

More ›