keeda

20 ms·

271.
by keeda 4mo ago
Here's a fun query to try with Google AI overview: "How many of the top 20 billionaires in the world have companies that have NOT been in legal trouble for anti-worker, anti-consumer or anti-competitive practices." The answer
272.
by keeda 4mo ago
The question needs to be tweaked a little: it's not just human vs LLM, it's human vs human + LLM, which makes the calculations easier (and more correct because LLMs don't currently operate independently.) I've run the na
273.
by keeda 4mo ago
Does it have to be a profit motive, or any motive would do?
274.
by keeda 4mo ago
> 3. The problem seems to be highly correlated with the increased financialization of our economies and the advent of a relatively recent set of so-called best practices about how companies should be built, run, and governed. I have th
275.
by keeda 4mo ago
It was the very first interaction with ChatGPT ever for me. I had dabbled some in NLP many years back, especially looking into the state of the art for summarization, and absolutely knew that we were at least half a century away from an
276.
by keeda 4mo ago
Oh, not just them -- Satya, Jensen and I are all on a first name basis. They just don't know it yet ;-)
277.
by keeda 4mo ago
Meat flapping at each other, no less. /shudder
278.
by keeda 4mo ago
So the more rigorous studies about AI-assisted coding productivity addressed this by keeping in place all other software development processes, including the same code review and quality standards, and only measuring throughput (PRs, LoC) b
279.
by keeda 4mo ago
There are other indications, however, like Anthropic paying through the nose for compute just months after Dario told Dwarkesh how hard it is to predict demand, or ChatGPT and Codex not quite having the same issues after Altman spent much-p
280.
by keeda 4mo ago
Right, what I meant is the underlying issue is the same, but the large amount of data along with the number of potentially conflicting and reinforcing biases going into LLMs make it hard to categorize or quantify risks. Like previously it w
281.
by keeda 4mo ago
I think we are talking about two different trends that have similar symptoms. I do agree there has been a noticeable anti-intellectual trend for a long time, especially in the West. (See also: Grade Inflation.) But that is separate from the
282.
by keeda 4mo ago
I get the sense a lot of the warnings about LLMs were based heavily on known risks of Machine Learning at the time (which those references are all examples of.) That was because the data was relatively narrow (e.g. hiring data.) However the
283.
by keeda 4mo ago
I'm sympathetic to the argument that AI is built without recompense on the work of all humanity, but anyone else think that's a lot of words to say "government backstop"? ;-) Personally, I'm very bullish on AI but i
284.
by keeda 4mo ago
> Many of them can no longer sit quietly for even 30 minutes just thinking on their own Plummeting attention spans has been a trend for much, much longer than LLMs and is more the result of constant digital interruptions and these day
285.
by keeda 4mo ago
Hmm, I would say the technology to durably preserve human knowledge is actually very recent. Until just a few decades ago, other methods have been extremely lossy. Even today, with triple-redundant, geographically-distributed data replicas
286.
by keeda 4mo ago
Have they had the exact same experiences and influences all the way to that point in time?
287.
by keeda 4mo ago
I've thoroughly enjoyed the couple of short stories of his I've read, so this was a highly disappointing read. Firstly, many of the technical arguments are of the "stochastic parrots" variety, which almost nobody really
288.
by keeda 4mo ago
Hmm, I doubt that would be an issue for a couple of reasons: 1. We have infinitely more durable means of storing, indexing and retrieving knowledge (and now potentially even means of programmatically reasoning about that knowledge!) We can
289.
by keeda 4mo ago
> Now, with AI being pushed on everybody even becoming something like a plumber is becoming a distant gap for the next generation. Yes, but this also has a silver lining: nobody will need to do the rote, toily, plumbingeque, but still
290.
by keeda 4mo ago
Hmm, despite the 13 years head start, turns out in terms of absolute and PPP GDP India and China seemed to have been neck-to-neck until ~1990, after which China really took off. https://data.worldbank.org/indicator/NY.G
291.
by keeda 4mo ago
Wait, where are you seeing the link to HAI? TFA mentions something called "liftlab" which seems to be something under Stanford Law School and separate from HAI. The study has more than a dozen authors from as many different univer
292.
by keeda 4mo ago
I'd say grade inflation is the clearest evidence. It has been a widespread, long-simmering issue; I've heard teachers bemoan the phenomenon in at least 3 different countries in as many decades, and each of them cited the ire of th
293.
by keeda 4mo ago
> ... but they have continued year over year for the past 3 years failed to deliver in AI. The vibes may not be with them but enterprise adoption seems to be: https://www.a16z.news/p/leaders-gainers-and-unexpected-
294.
by keeda 4mo ago
> While your linked post was explicitly about ads, this line wasn't and is just baseless I mean, the ad business is still by far the largest chunk of their whole business, which is why I think their overall revenues, profits and d
295.
by keeda 4mo ago
Maybe, but Microsoft, through their partnership with OpenAI, is already involved in major copyright lawsuits. That is probably a driving force for this move, actually... I doubt they would want to tempt fate while those lawsuits are on-go
296.
by keeda 4mo ago
Hah, I have a prototype of the same idea on my backburner! Excited to see this, though I don't understand some of their design choices. Will need to check out more closely.
297.
by keeda 4mo ago
> Second, clean data. MAI-Thinking-1 was trained on clean and appropriately licensed data, with AI-generated content excluded from pre-training. This matters for quality, provenance, and control. If we cannot account for what shaped a m
298.
by keeda 4mo ago
Right, but my theory is that the ad business, still the biggest chunk (75%+) of their revenue, is extremely hyper-optimized for the current search journey-based UX and enjoys a monopoly + auction rigging premium (e.g. Project Bernanke) prim
299.
by keeda 5mo ago
I agree that all of today's CEOs have learned from history and are paranoid about disruption, and I agree that Google is pivoting effectively and will even thrive in the AI era, given their technical and distribution advantages... but
300.
by keeda 5mo ago
As a layman external observer, this seems a bit rushed to me? I know there's a race amongst the frontier AI labs, but I don't quite understand the rush (apparently there is still a lot of money to go around), so as some comments i

More ›