schleck8

Why do you cite him from 2024 when he spoke about it much more recently? This is a quickly moving field after all

“I think we’re going to see AI get even better,” Hinton replied. “It’s already extremely good. We’re going to see it having the capabilities to replace many, many jobs. It’s already able to replace jobs in call centers, but it’s going to be able to replace many other jobs.”

https://fortune.com/2025/12/28/geoffrey-hinton-godfather-of-...

OpenAI's Orion (GPT 5/Next) is partially trained on synthetic data generated with a large version of o1. Which means if that works the data scarcity issue is more or less solved.

Can't wait to give them access to everything I do on the daily by wearing their AR glasses.

The average person doesn't know what a DNS is and isn't willing to pay for a VPN. Society isn't a tech bubble.

The US is in no position to critique other countries' checks and balances right now with how the Supreme Court works

Okay, but this isn't about privacy since Telegram isn't even end to end encrypted by default where it matters.

What happens then? They leave a vacuum and then what? Noone fills that vacuum? Assuming there is zero competence in the EU, which is highly unlikely since both the best image generation model right now and very respectable open source llms are from the EU, and on top of this several countries in Europe have exceptional tech talent (especially in the East), the Chinese would jump in immidiately.

This is about identity theft like selling credit card details, meth and child porn (telegram deletes 60 thousand such groups. Per month. It's a shitpile that attracts flies, evidently)

Hey, get away with your facts and logic. It distracts from OP's agenda.

It's not targeted at farmers clearly. The page literally says the xl version is for a family of four

The rotary tool features a 24 volt DC motor, interchangeable implements, and an adjustable motor angle allowing FarmBot to perform light duty weed whacking, soil surface milling, and drilling operations.

Schnell is definitely worse in quality, although still impressive (it gets text right). Dev is the really good one that arguably outperforms the new Midjourney 6.1

This makes no sense. Benchmarking code is easier than natural language and Mistral has separate benchmarks for prominent languages.

I think I've only ever seen a single incorrect answer from Perplexity and I've probably made a thousand searches so far. It's very reliable

Great explanation for the type of person to think the Gaza conflict started on October 7th 2023.

The Georgian allegations of a Russian invasion were supported, inter alia, by claims of illegal entry into South Ossetia of a large number of Russian troops and armour, prior to the commencement of the Georgian operation. According to Georgian answers to the Mission´s questions, the process of building-up of Russian forces in South Ossetia had started in early July 2008, continued in the course of August and included troops and medical personnel, tents, armoured vehicles, tanks, self-propelled artillery and artillery guns

(From your source)

Tskhinvali is in Georgia by the way. I think that says it all.

Read this: https://en.wikipedia.org/wiki/Russo-Georgian_War

For context, Russia invaded Georgia in 2008 and still occupies a fifth of the country today.

https://youtu.be/quipA4hHCV4

Oh, absolutely. Never heard of the LinkedIn scraping lawsuit before? They are totally willing to throw their lawyers at you for using the platform in a way it's not intended to be used in

Out of the links in the footer:

- Painkiller Ideas and Replyguy both use Azure DNS and Microsoft hosting which means the footer probably contains the creators' other projects

- while stealth.marketing is hosted very differently, it uses partially the same and partially very similar CSS

- stealth.marketing says it's "powered by reddit" with the reddit logo which I'm sure the Reddit PR department will absolutely hate when finding out, given they ruin the platform (probably not even using the API)

Looks like there's an easy way to make this all collapse

Does this break the OpenAI terms of service? Or whatever other API they query?

They are gonna get sued if they try this on LinkedIn or Tiktok for sure. Circumventing those captchas (everyone who's tried scraping Tiktok before knows what I'm talking about, even in headed environments) is almost certainly illegal.

I think so too. Talking about monopoly and then participating in advancing a de-facto monopoly (Chromium on Android and Chromium on Windows)

I see third party antivirus all the time. Specifically I've seen Setinel One and Sophos in enterprise applications like hospitals. I'm pretty sure the enterprise version of Defender costs money so you might as well use an alternative that better fits your admins' needs etc.

The image creation isn't Llama 3, it's not multimodal yet. And the web search is Google and Bing API calls so just use Copilot or Perplexity.

Zuck equated the current point in AI to iOS vs Android and MacOS vs Windows. He thinks there will be an open ecosystem and a closed one coexisting if I got that correctly, and thinks he can make the former.

Depends on your size threshhold. For anything beyond 100 bn in market cap certainly. There is some relatively large companies with a similar flair though, like Cohere and obviously Mistral.

You absolutely can since it has a size advantage either way. MoE means the expert model performs better BECAUSE of the overall model size.

Just use the Replicate demo instead, you can even alter the inference parameters

https://llama3.replicate.dev/

Or run a jupyter notebook from Unsloth on Colab

https://huggingface.co/unsloth/llama-3-8b-bnb-4bit

the EU laws are getting too draconian

You also said that when Meta delayed the Threads release by a few weeks in the EU. I recommend reading the princess on a pea fairytale since you seem to be quite sheltered, using the term draconian as liberally.

Neglected to include comparisons against GPT-4-Turbo or Claude Opus, so I guess it's far from being a frontier model

Yeah, almost like comparing a 70b model with a 1.8 trillion parameter model doesn't make any sense when you have a 400b model pending release.

100 % job placement rate 1 % of the time