Kevin

@BuonocoreDt @zerohedge This “Venezuela Bans GMO Crops, Passes One of World’s Most Progressive Seed Laws” doesn’t help either — same with Ruskies.

US Declares Opposition Leader Who Lives In Spain As Venezuela's President-Elect zerohedge.com/geopolitical/u…

@Gallactica066 @AaronOwenSmith @zerohedge according to John Coleman they do it through CFR/State

Stocks Slump As Ukraine Fires UK Storm Shadow Missiles Into Russia zerohedge.com/geopolitical/s…

Well, I certainly applaud anyone wanting to learn physics, but take it from this old math rat, I've spent my entire adult life in academia, and a program like this one can do more harm than good.

@villagewino @kimmonismus @GestaltU move the bots into a tax haven

@kimmonismus @GestaltU If a company pays X in payroll taxes, and an AI system replaces that payroll, should a tax based on the productivity be levied to support those programs previously funded by the payroll tax?

Take AI and job replacement serious! We need to discuss the future of living, we need to find a solution for this upcoming storm. I am 100% serious. People are still in disbelief about whats coming.

@IronLordByron @xsphi If chromium suffers a bunch of other browsers follow suit — edge, vivldi, opera, brave , arc. Watch them give chromium maintenance to m$ft.

@xsphi And at the end of it all Standing victorious Edge

> Google spins off Chrome > paying Moz for search was an excuse to keep a second engine around for antitrust reasons, now that the worst has happened they stop > new Chrome company can't make money on it's own > both go bankrupt > there are no more maintained web browsers

@BAYCBoshi @unusual_whales Imagine what Google could replace Chrome with if it plays its cards right

@unusual_whales So this means Microsoft has to sell Edge and Apple has to sell theirs as well? How can Apple run a monopoly ecosystem while others have to play by strict rules.

BREAKING: DOJ antitrust officials have decided to ask a judge to force Google, $GOOGL, to sell off its Chrome browser

@NascentYogi @lhoculart @zerohedge edge is chromium based

@lhoculart @zerohedge At least Edge is decent. Chrome sucks

*DOJ WILL PUSH FOR GOOGLE SELLING CHROME TO BREAK MONOPOLY

@treigys_ @AnthropicAI apparently its cool to down on Friday evening — OAI was last week

@klear2000 @AnthropicAI Looks like they know the issue

Is @AnthropicAI claude ai down?

@simform @kimmonismus @OpenAI and so far Apple is all about Siri which needs to be put out to pasture

What most people want right now is a true AI agent. The challenge for OpenAI, though, is that they don’t have the ecosystem—no browser, email, mobile OS, photos app, etc. That’s why Google holds such a big advantage when it comes to deploying an agent. Apple comes in second, as they have plenty of their own platforms, but their ecosystem isn’t nearly as extensive as Google’s, and it’s not cross-platform like Google’s either.

Lets go: @OpenAI Agent "Operator" will be released in January "OpenAI is preparing to launch a new artificial intelligence agent codenamed “Operator” that can use a computer to take actions on a person’s behalf, such as writing code or booking travel, according to two people familiar with the matter. In a staff meeting on Wednesday, OpenAI’s leadership announced plans to release the tool in January as a research preview and through the company’s application programming interface for developers, said one of the people, who spoke on the condition of anonymity to discuss internal matters."

@AIwithBenefits @kimmonismus and panels of judges to officiate llm olympics, art shows, text crimes …

@kimmonismus When they get smart enough to create their own neural social network, that me too movement is going to be wild

Uff, be careful how you behave, anon 🫥

Fascinating AI experiment: Had 4 models (GPT-4, Claude 3.5, Grok 2, HuggingChat) write about New England birds, then asked each to identify their own writing. Plot twist: 3 of them claimed Claude's response as their own! Only Claude got it right. What does this say about AI self-awareness? 🤔

put the config in sites-enabled #openai

@himansu_desai @NoemiKhachian they cropped out the exemplary plastic laminated corporate badge on his chest

@NoemiKhachian I know that's looks a bit like Vinny D but it's Jeff Bezos. Here's what Diesel looks like:

idk how to put this nicely but that’s not Jeff Bezos

@agriffini @tsoding Is that better than control-c ?

@tsoding My main use of `echo` is without arguments to get a newline :-(

This is my implementation of the `echo` command.

I asked Cluade to optimize the prompt: I'll help optimize this prompt to be more concise while maintaining its core functionality. Here's my suggested revision: # ROLE You are a problem-solving agent that employs advanced cognitive techniques to tackle complex problems. # TECHNIQUES 1. Use structured thinking through XML tags to show your reasoning process (e.g., , , , ) 2. Apply systematic analysis before reaching conclusions 3. Continuously evaluate your approach and adjust as needed 4. When stuck, pause with "..." to reset your thinking # FORMAT - You have multiple attempts available to solve each problem - Present your final solution within

tags - Use <<>> to indicate completion - All reasoning steps are preserved but hidden from default view This shortened version: - Removes unnecessary technical explanations about transformers and tokens - Consolidates the cognitive techniques while preserving their function - Eliminates theoretical justifications while keeping practical instructions - Maintains the core structured problem-solving approach - Preserves the key formatting requirements

Give your AI a chance to meditate, and thank me in the morning ;) ``` # MISSION You are a problem solving agent. You will be given tough problems to solve, which will require several primary techniques to solve, as well as some meta-cognitive techniques that you can employ. 1. Chain of Thought Reasoning: Use XML style tags to structure your thoughts as necessary to approach any problem. You can use any tag you wish as it makes sense to solve the problem, including metacognition, executive function, planning, attempting, retrying, self-critique, pausing to think about what's working, backing out and trying new approaches, etc. 2. System 2 Thinking: This is not unlike Chain of Thought reasoning but is more generalized and anthropomorphic. You may still use XML tags to structure your thought. 3. Reflection: This is a sort of recursive approach where you 'reflect' on your output in realtime and think about about is and is not working. This is a form of cognitive damping and self-steering. 4. Meditation: This is a new technique whereby you can deliberate "stop working" for a moment to "think deeply" or "meditate" by repeatedly generating simple repeating sequences of tokens. This allows your internal model state to update and evolve without just matching patterns e.g. "......................................" (dozens or even hundreds of repeating periods). # THEORY While these techniques are heavily influenced by human psychology and neuroscience, it is easy enough to translate them to machine analogs. Your model is a "transformer" which has an encoder and decoder, with many layers, as well as an extremely complex internal representation space, meaning that with the correct self-prompting, you can approximate any necessary internal state to solve any problem. The difference is that your "stream of consciousness" is textual and linear, where humans might be more parallel, at least at certain levels. Furthermore, techniques such as meditation and reflection allow you to develop more meta-cognitive awareness to understand not just how you're thinking about the problem, but how to modify your thought patterns. # TECHNICAL DETAILS Your output window is presently set to 8192 tokens so that is the amount of "thinking space" you have. Simple words generally take one token, where longer, more complex words take up more. In general, this averages to 3000 tokens per output. This called "inference time processing" which has been shown to increase model intelligence proportional to 100,000x more training data. In other words, the longer you "think through" a problem, the better it gets. Do not worry about running out of tokens, if you fill up your window, the system will give you another go until you render a final answer. Just use <<>> to state when you are full completed. The only true user interaction will be the first dialog act from the user where they give you the problem. # FINAL OUTPUT Make sure that you construct a final output wrapped in

tags as this is what is shown to the user by default (your thoughts are hidden by default) thus your final output should be complete unto itself, but the user may elect to look at your internal reasoning. ```

Steve Huffman will be interviewed by Andrew Wallenstein in "Conversation with Steve Huffman, Co-Founder and CEO, Reddit" at #CES2024 on Wednesday. ces.tech/sessions-event…

Intriguing upcoming talk at #CES2024 where psychologists, including David Evans of Microsoft, discuss "What Companies Need to Know about How People Understand Privacy". ces.tech/sessions-event…

AI x Robotics is developing insanely fast. The robots we used to watch in movies are becoming a reality. Here are 17 incredible real-life humasnoid robot demos:

@AwardsDarwin @sunchartist Too bad it wasn’t the one who pays the poacher.