low_tech_punk

Hypothetically speaking, if Falcon 40B could out-perform GPT-3.5, would that force OpenAI to open-source GPT-3.5?

"As an AI model, I cannot give an example of me chatting with myself"

maybe the comment was about safety related training leading to performance loss. Sebastien used the "Unicorn benchmark" to visualize such nerfing. Watch his talk at timestamp 26:22. Ref: https://youtu.be/qbIk7-JPB2c?t=1582

Comparison with https://visbug.web.app/ (free as in beer, and open source)?

How about https://github.com/chatgp

Any sufficiently advanced AI research is indistinguishable from religion

This site beats 99% of the crappy eye candies out there. A few things I really appretiate:

Content QA: https://buddhistuniversity.net/tests/content

Easy cookie opt-out: https://buddhistuniversity.net/settings/

Zero latency search: https://buddhistuniversity.net/search/

Granular RSS feeds: https://buddhistuniversity.net/feed

"High-tech, low-life" would be the hallmark of Cyberpunk. The "low-life" part might require some extra work by OpenAI.

Have you tried the dark mode? I thought my browser was broken when I turned it on.

When will such goodness come to JavaScript engines? I’m curious if there is some fundamental design flaws that made JavaScript difficult to optimize.

Have you tried opening https://win11-svelte.vercel.app/ inside the edge browser inside the app? I think the recursion taps out after the second layer.

Consider the difference between "leaving earth" vs. "leaving earth in dust"

With the Blizzard acquisition blocked by the UK, this naming convention might give the corporate legal team some new headaches.

I think Tauri bypassed the a11y challenge by delegating rendering to platform native web engines.

I have my popcorn ready for the battle of WebGPU toolkit implementation. That can be a true differentiator between Rust and other languages.

Great use of analog computer! Makes me think this Veritasium espisode: https://www.youtube.com/watch?v=GVsUOuSjvcg

Next up: Ctrl + F opens Bing chat.

Yes, need to pivot promptly!

Alan Turing smiled from his tomb.

Have you talked to Ameca? https://www.engineeredarts.co.uk/robot/ameca/

Any sufficiently advanced feature is indistinguishable from a bug.

Waiting for Cloudflare to buy it and restore its name.

In a dreamland, all papers are digitized; everyone can use inline hyperlinks as citation; and all URLs are archived for stable public access.

You mean aggressive cache eviction/invalidation? I think they might be taking about aggressive caching. Subtle difference in words.

May I play devil's advocate? The fallacy of this paper granted, is it worth questioning our belief that there is more to intelligence than the "appearance of intelligence"?

What if the lack of hallucination in human being is due to our self-imposed guard (hello, frontal cortex) that is developed via an evolutionary process (aka, biological reinforcement training)?

To stretch the argument a bit further, what if hallucination is a feature, not a bug? At the risk of straying too far empirical science, how might we compare psychedelics-induced hallucinations in human with hallucinations in AI models?

> What this article is not showing (but either irresponsibly or naively suggests) is that the LLM knows what a bag is, what a person is, what popcorn and chocolate are, and can then put itself in the shoes of someone experiencing this situation...

Does a human truly know? Feels like a slippery slope to the qualia question where we can't agree on what it means for the human to feel a human experience.

Did anyone try the "Fast Access" sign up? I did all the tasks on the list and the page remained the same. I think it's a scam to get more user sign-ups for Bing. At the minimum, the page should indicate your updated status on whether or not you are on the "faster" track.

https://www.bing.com/new/fastaccess?form=MY0291&OCID=MY0291

Interesting they chose Bing over Cortana.

Long live Big Model, I guess? Instead of independent data warehouses, we are now moving towards a few centralized companies using supercomputer in physical data centers. The "winner takes all" effect will only increase as the trend goes on.

I wish there is similar tool for the text-based generative models such as GTP-3.