Not sure I agree with his tests, but I agree with the headline, I recently had cursor launch into seemingly endless loops of grepping and `cd` and `ls` files. This was in multiple new convos. I think it's they're trying to do to much, for two many "vibe coders", and the lighter weight version that did less were easier to steer to meet your architecture and needs.
I kind of get it, except youtube... which has much more educational, news, and long form content. Also also forcing face/age verification sounds ripe with issues.
It has some educational content, most of it is brain rot like everywhere else though. Open a brand new youtube account and check out what's being pushed by default, you either get room temperature IQ political analysts or "shorts" with softcore porn thumbnails to bait people for a click
Open a new YT account then feed it with [1] for few hours at least then you will unleash the full power of Youtube... unless you missclick even once into some popular blog typically they very clearly aimed at low-IQ people which accidentably might be your kid or somebody else like you know who I mean. But to prevent that slippery slope at least partly, just increase the feeding time of your YT account with the best requests possible which are carefully stacked at [1].
Sure, but then again that's not how most people use youtube in real life. You can trick it into temporarily not being a slop provider but it's a constant battle, and not one a 10-15 years old will engage in, which is to be expected, kids are kids and can't fight multi billion dollar companies hiring the top behavioural scientist of our era to create the most addictive ad delivery mechanism possible
dear god, I wonder what the accuracy rate on these predictions will be "Does this work against the new smart-mattresses? Mine refuses to soften up unless I watch a 30-second ad for insurance." <https://sw.vtom.net/hn35/pages/90098444.html>
Wow, that is incredible. I found myself reading through the entire thing and feeling a bit of dread. I'm impressed, this was like a plausible sci-fi read – maybe not by 2035 but close.
They don't, akamai has had several outages as well jsut no one notices. Akamai is way way smaller than cloudflare, 20% of internet traffic passes through CF networks, not sure it's even measurable on Akamai.
Quickly Googling about, a commonly repeated figure is that Akamai served 15% - 30% of Internet traffic in the late 2010's. They probably have less of the market today due to others growing, but they're not a minnow.
2024 revenue figures were $1.669 billion for Cloudflare, and $3.99 billion for Akamai, per Wikipedia.
> 20% of internet traffic passes through CF networks
That does not sound right to me. “20 percent of websites” does not mean “20 percent of traffic.”.
There is no public write-up from Cloudflare that proves “we handle 20% of all Internet traffic.” Cloudflare reports around 295,000 paying customers and more than 30 million Internet properties (20% of the web). So most of their users are on the free plan.
agreed, cloudflare says 20% of all websties, which makes sense, digging into it the "20% of all web traffic" seems to have been like a game of telephone in media, social media, and/or AI.
I can't believe they only have 295,000 paying customer, that puts me in a small minority. lol
8T is the high-end of the McKinsey estimate that is 4-8T, by 20230. That includes non-AI data-centre IT, AI data-centre, and power infrastructure build out, also including real estate for data centres.
Not all of it would be debt. Google, Meta, Microsoft and AWS have massive profit to fund their build outs. Power infrastructure will be funded by govts and tax dollars.
There is mounting evidence that even places like Meta are increasing their leverage (debt load) to fund this scale out. They're also starting to do accounting tricks like longer depreciation for assets which degrade quickly, such as GPUs (all the big clouds increasing their hardware depreciation from 2-3-4 years to 6), which makes their financial numbers look better but might not mean that all that hardware is still usable at production levels 6 years from now.
They're all starting to strain under all this AI pressure, even with their mega profits.
"But we bought too many GPUs! We spent billions on infrastructure! They have to be put to work!"
right... none of them are saying that. They could probably use more GPUs considering the price of GPUs and memory are skyrocketing and the supply chain can't keep up. It's about experimentation, they need real users and real feedback to know how to improve the current generation of models, and figure out how to monetise them.
The comment I’m replying to is an example of one hallucination. Specifically that AI is being pushed because companies have too many GPUs when in reality they have too little. There are several other hallucinations in this thread.
This will need a separate blog post But when you give something for free, then you will run out of that resource. So yes, companies have too little GPUs to give their services for free, but too many GPUs for their paid services.
Just to be clear, are you asserting that every opinion in this thread that you don't agree with is due to the poster hallucinating, or only specific ones?
Do you have any evidence or well established theory to back up this rather extraordinary claim?
Because if you are honestly positing that numerous people around the world are literally hallucinating despise (statistically) not being under medical supervision, presumably continuing to drive, work, and make decisions, that would be a pretty urgent global health phenomenon that you really should be chasing up.
And at some point, the authorities best placed to deal with this hitherto unseen mass incapacitation might reasonably ask: what are the chances that multiple unrelated people around the world are experiencing such localised, hugely specific breaks from reality causing them to express reasonably common opinions on an internet forum, rather than the inconsistency being on the end of this one person who doesn't agree with them?
That's a fair point! We were aware of onnx, but felt it was okay since they are very different products so we felt that there wouldn't be too much confusion (people generally know which onyx/onnx they are looking for).
You're almost literally in the same ecosystem, it's not like one is a Chat UI for LLMs and the other a super market, but a ecosystem of open source machine learning software, libraries and tools. That the pronunciation is identical makes it untenable, you really need to reconsider the name, discussions in person will get confusing.
reply