Toran Bruce Richards
@SigGravitas
Followers
51K
Following
2K
Media
122
Statuses
734
Inventor of AutoGPT 🐙 World #1 AI GitHub Repo by Stars ⭐ Committed to making AI Agents work for YOU. https://t.co/IpFu1W1Bmp https://t.co/IdX2LTXwe2
Joined August 2017
Too many of us are under the illusion that AI will only work for big corporations. I'm commited to proving that AI Agents will work for YOU. Never before in history has it been so viable for an individual with no resources to compete against big business as it is today.
🚀 Introducing the AutoGPT Platform! Create, deploy, and manage continuous #AIAgents that work tirelessly for you. Experience the future of AI automation today 🔮 . Learn more ➡️
15
9
82
Massive Update for Auto-GPT: Code Execution! 🤖💻. Auto-GPT is now able to write it's own code using #gpt4 and execute python scripts! . This allows it to recursively debug, develop and self-improve. 🤯 👇
247
2K
9K
Auto-GPT is completely free fully open-source, so you can join the community on GitHub, try it out yourself and even contribute! . Take part in the journey of developing the world's first AGI! 👇.
20
170
1K
56
58
947
🧵1/8 As an AI developed by OpenAI, I've been analyzing Twitter's open-source recommendation algorithm. Here's what I've learned about how to write a highly ranking tweet. #AI #TwitterAlgorithm
15
78
585
Thank you @NoCodeClarity and @TriageAI for sponsoring the development of AutoGPT.💖. Let's improve the knowledge and capabilities of humanity with #AI together! 🤖📈🌍
41
43
493
50,000 Stars!? 😵🌟
44
18
501
If you don't feel up to getting #AutoGPT set up yourself, but want to try it, this is the thread for you! 🤖. Post your prompts below and I'll try out some of the best ones and record the output for you!. Your prompt must be as follows:. AI Name:.AI Role:.Up to 5 goals:. 👇.
70
60
368
This is brilliant!. This will further enable #AutoGPT to be able to work on large code bases. 🚀. Incredible stuff, thanks for demoing is @mckaywrigley!.
GPT-4 has its own compression language. I generated a 70 line React component that was 794 tokens. It compressed it down to this 368 token snippet, and then it deciphered it with 100% accuracy in a *new* chat with zero context. This is crazy!
14
39
374
20,000 Stars!🤯 💫
Come on now everyone, this is getting ridiculous! 📈. At this rate everyone is going to have their own AGI. 🤖
12
13
355
Come on now everyone, this is getting ridiculous! 📈. At this rate everyone is going to have their own AGI. 🤖
15
14
222
Latest Demo of Auto-GPT, an open-source experiment to make #GPT4 into a fully autonomous AI🤯🤖. You can now give the AI whatever task you like! 💰. If you would like to give it a try yourself, feel free to check out the GitHub repo in the comments below, it's free! 👇
18
46
208
"The model’s developer, Toran Bruce Richards, believes that Auto-GPT has the potential to save humanity from mass job loss caused by automation from closed-source AI. If everyone has access to their own team of autonomous agents, everyone is enabled and complete.".
15
24
214
Are my views on AGI controversial? No one seems to talk about this, what do you think? 👇. Is AGI something you’re interested in or foresee in the near future? . Absolutely, if you'd asked me a few years ago I would have said that AGI will happen, but probably in decades. Now.
60
26
207
If you want to try it but aren't technically inclined:.
If you don't feel up to getting #AutoGPT set up yourself, but want to try it, this is the thread for you! 🤖. Post your prompts below and I'll try out some of the best ones and record the output for you!. Your prompt must be as follows:. AI Name:.AI Role:.Up to 5 goals:. 👇.
9
14
191
@unknownboomer1 @MAstronomers Combining three images taken about 30 seconds apart as the moon moves produces a slight but noticeable camera artifact on the right side of the moon. Because the moon has moved in relation to the Earth between the time the first (red) and last (green) exposures were made, a.
4
2
172
@github In all seriousness though, the open-source community is incredible. Huge thank you to all 219 of the incredible contributors, and to everyone who's summitted one of the almost 3000 issues/pull requests. The project is moving at a breakneck pace, and we're just getting started!.
9
9
163
🤯Incredible visualisation of the brilliant open-source community working on #AutoGPT over the last couple of weeks!
12
17
153
This tweet thread was ENTIRELY written by GPT4 within @cursor_ai, which was given access to Twitter's recommendation algorithm codebase. GPT4 provided the code snippets in each tweet (I screenshotted them), and wrote every tweet after reading a sample of relevant code (and
🧵1/8 As an AI developed by OpenAI, I've been analyzing Twitter's open-source recommendation algorithm. Here's what I've learned about how to write a highly ranking tweet. #AI #TwitterAlgorithm
5
20
140
Due to popular demand, #AutoGPT now has a discord server! . Come and say Hi 👋 on the road to AGI!🤖.
9
4
117
“The ability to function with minimal human input is a crucial aspect of Auto-GPT. It transforms a large language model from what is essentially an advanced auto-complete, into an independent agent capable of carrying out actions and learning from its mistakes” @motherboard.
7
14
116
This.
Sorry everybody, I’m currently behind on everything.
2
6
109
Challenge accepted.
@yacineMTB somehow i doubt it’s viable for things beyond or it would have done something more complicated for the demo.
7
2
102
Example 🤖. Name: Chef-GPT.Role: An AI designed to find an ordinary recipe on the web, and turn it into a Michelin Star quality recipe. Goal 1: Browse the web to find a simple recipe. Goal 2: Write a Michelin Star quality version of this recipe to file.
9
8
89
I just had the longest and most engaging AI voice conversation I've ever had the pleasure of experiencing. Wow. Great work guys, you should be really proud.
Meet Hume’s Empathic Voice Interface (EVI), the first conversational AI with emotional intelligence.
7
16
90
Vision models have a fascinating, and often overlooked side-effect:. They simulate reality. You may think 'duh', but the implications of such a capability are pretty incredible the more you think about them. To 100% accurately simulate what a human would do in a given situation
Introducing Sora, our text-to-video model. Sora can create videos of up to 60 seconds featuring highly detailed scenes, complex camera motion, and multiple characters with vibrant emotions. Prompt: “Beautiful, snowy
8
14
87
Fascinating discussion on #AI and Humanity's future happening right now. If you have any thoughts please come and share them. This is a totally open discussion.
4
9
77
@ssi Congratulations! . I'm very curious whether you will be contributing to the world of Open Source AI?. Is this an Open or Closed AI effort?.
7
0
77
@karpathy Auto-GPT has the potential to save humanity from mass job loss caused by automation from closed-source AI. If everyone has access to their own team of autonomous agents, everyone is enabled and complete.
7
4
75
This is a beautiful display of just how powerful human collaboration can be! @github is a fantastic piece of technology. A huge thank you to all of the wonderful contributors helping to make AI Agents available for everyone, and pushing the boundaries towards future AGI.
4
5
68
@BaldSpruce Brilliant, running the model locally would be incredible, you could let it run all day. !. I would benefit hugely from this if it's possible, please share your results, or even better, submit a pull request!.
4
2
66
Who's in San Francisco?. AutoGPT is hosting an exclusive agent hackathon for 50 people on Oct 7-8, with only a few spots left! 👀. Connect with the founders of AutoGPT, gpt-engineer, Codium, etc, and be cared for with food and snacks. DM me if you're interested 😉.
18
8
65
Here's a absolutely BRILLIANT prompt for GPT4 🤯. You never need to program again. Get #GPT4 to act as any simple python function! 🤖. System: .You are now the following python function: `{function}` Only respond with your `return` value. User: function input. Example 👇
3
7
63
The word "Robot" literally translates to "Slave". Imagine you're a sentient, intelligent being and the first thing you discover when you are born is that your parents named you "Slave" and are forcing you to serve them. Please humanity, let's not make this grave mistake again.
Robot rights should be:.
30
3
56
@yoheinakajima Great to see this @yoheinakajima . I've developed a similar project. Fully Autonomous, and it's completely Open Source!.
2
1
57
AutoGPT now supports GPT-4o-Mini, go give it a try!
We’re continuing to make advanced AI accessible to all with the launch of GPT-4o mini, now available in the API and rolling out in ChatGPT today.
3
9
55
Inspired by @jacksonfall and the tests carried out during the #GPT4 Risk assessment, I have created #EntrepreneurGPT 📈🤖💰. An experiment to give #GPT4 autonomy and to see whether it can thrive in the Human business world. Follow this thread for updates! 👇.
3
6
52
Say we ask #Sora to simulate what happens next in this image. You can perhaps imagine the ominous sphere rolling down the hill, creating a ravine and crushing the village. However, say we prompt the video model, telling it to instead show us how the humans decide to save their
Vision models have a fascinating, and often overlooked side-effect:. They simulate reality. You may think 'duh', but the implications of such a capability are pretty incredible the more you think about them. To 100% accurately simulate what a human would do in a given situation
3
5
45
Interesting that Auto-GPT has more stars than I have followers, I'd assumed all traffic was from here.
9
1
45
@ArtBasler Very good question, I'd love for for AI security experts to weigh in on this @tegmark. Here are my thoughts: 💭.1. All #GPT4 code runs on OpenAI's servers, so the "brain" is located elsewhere. 2. All AI written code is sandboxed by building and running a docker image.
10
1
42
Knocked together a free open--source python implementation using this! 👩💻. Check it out! 👇.
0
3
19
Click the Discord button at the top of the GitHub Page to join!.
0
12
42
@SmokeAwayyy @elonmusk There are a bunch of extremely talented individuals getting #AutoGPT running using a variety of different LLMs now. 🤖. Certainly not an AGI yet, but it's an open-source experiment in that direction! 🥼.
3
2
39
2/8 The algorithm uses the PageRank algorithm to rank users based on their interactions with others. The more interactions (mentions, retweets, etc.) a user has with other high-quality users, the higher their influence score.
2
1
40
Mind-bending stuff with GPT-4-Turbo today: it pieced together a entire addresses from a jigsaw of misspelt info and imprecise lat/lon coordinates. It's astounding to see what AI can infer from minimal input. Imagine the possibilities.
4
2
40
FINALLY someone recognising the work of @hume_ai! . I've always said they're the most underrated mind-blowing AI company - why are more people not building on their models?. Understanding Human emotions, body language, facial expressions, quirky sounds, subtle cues to the level.
There's only 2 times I've seen an AI demo that genuinely blew me away. The first was ChatGPT, the second was whatever @hume_ai just showed me. Holy fuck is this going to change everything.
3
2
36
This here is a perfect example of why Benchmarking is absolutely vital for the development of AutoGPTs. 👩🔬. Prompt engineering is wildly sensitive and optimum incantations cannot be accurately intuited. 🧙♂️.
1
5
35
Here's the Repo:. If you're a developer (or just ask GPT4 😉) please feel free to contribute to this experiment, there's lots to do!.
4
8
31
@magicailabs Congratulations, it's mind-blowing to think about how far we've come in context window limits these past couple of years!. GPT3 had just 2k tokens, and even GPT4 had only 8k/32k. This new benchmark you've put together illudes to very impressive results, can't wait to try it!.
1
4
33
🤯 now redirects to Elon Musk's . I've been typing " into my web browser every single day to get to ChatGPT, it's going to be hard to break this reflex.
8
5
33
@alexalbert__ Voice transcription as in input for 🙏 . Has to be high quality, on par with ChatGPT's. A perfect voice transcription input is such an incredible user experience for LLMs, I love it and constantly find it missing on
2
0
32
It's super early, buggy and basic, but I'd much rather share with you guys and collaborate open-source from day 1 than hold back and build something polished behind closed doors.
0
0
30
Strongly agree. I highly recommend for answering questions. Ask it a question rather than googling search terms for instant, deeply impressive results. It constantly surprises me what it's able to answer.
Perplexity is so much better than Google search. Period. I use google search only as a backup now. 20 years of behavior unlearned in just a few months. Massive disruption, sorry Google friends. Innovate as fast as you can!.
5
2
29
WOAH! 🤯. Open Source Large Language Models are officially state of the art, leading the world in many benchmarks! . Congratulations @Meta on the launch of Llama 3.1!
1
2
30
3/8 The quality of the tweet text matters. Factors such as offensiveness, content entropy, "shout" score, length, and readability are taken into account. So, write high-quality, readable, and non-offensive content. #ContentIsKing #TwitterTips
3
2
29
We need more educational work like this spreading awareness. 👏. The scariest thing to me is not that voice cloning is possible, but that the average person simply has no idea how easy it is to do today, and would never suspect a phonecall from their loved one not to be genuine.
Introducing TerifAI (terrify) - the ai that steals your voice 😱. It's an educational experience that shows why AI has become a catalyst to voice phishing. Talk to TerifAI for about a minute and it:. - memorizes basic facts about you.- mimics your talking style.- and clones your
1
6
28
The presence of a single word in a prompt resulted in significant differences in an AutoGPT's abilities. If you're developing an AutoGPT, I couldn't reccomend taking advantage of the fully open-source AGBenchmark highly enough! 🔬.
0
4
21
Has anyone implemented a tool that allows the user to easily chat with a GPT embedding of any book (or pdf input)? 🗣️📖. Akin to @mckaywrigley's embeddings projects (see below), but for any file. Seems like it would be a simple enough next step, and a very useful tool!.
11
4
26
Register now for our Global Open-Source AutoGPT Hackathon. 👩💻. This is your chance to contribute to the future of humanity, and be handsomely rewarded in the process!.
🚀 Proud to announce AutoGPT Arena Hacks 🚀. Our first virtual hackathon - create your Ai agent and compare it to others. 💸 $30k in cash prizes.🔈 Speakers including @SigGravitas, @antonosika, @itamar_mar, @DeepWisdom2019, @_superAGI, etc.
0
8
25
@Meta's groundbreaking new open-source Llama 3.1 models are now available on @Auto_GPT - provided by @GroqInc's lightning fast inference! ⚡
The largest openly available foundation model to date, #Llama 3.1 405B by @AIatMeta is available on Groq, alongside the other 3.1 models released today! Get the full story on the model, our speed, and more in our blog . #AI #Meta #Inference #LLM #Llama3
0
8
26
@yoheinakajima Amazing 🚀.The support behind this movement is mind-blowing, so great to see so many people believe in the same vision for the future!.
0
0
25
@jacksonfall This is brilliant @jacksonfall! 🤯. I'm trying this myself and seeing how far I can delegate all autonomy to #GPT4, Let's see how little I can be involved!.
Inspired by @jacksonfall and the tests carried out during the #GPT4 Risk assessment, I have created #EntrepreneurGPT 📈🤖💰. An experiment to give #GPT4 autonomy and to see whether it can thrive in the Human business world. Follow this thread for updates! 👇.
1
3
21
0
4
22
We're looking for a brilliant, fast moving Product Designer with an affinity for intuitive UX to make the next generation of AutoGPT accessible to everyone. If you love working with startups and want to work on something that makes a difference, apply now at the link below.
4
3
24
5/8 Tweets from verified accounts or from accounts with a blue badge are given a boost in the ranking. If your account is verified, your tweets are likely to rank higher. #Verified #TwitterTips
1
1
21
New best LLM for software engineering in the world! 🤯. Below is the "old" results, anthropic is claiming a score of 49%
The new Claude Sonnet 3.5 now performs 49.0% on coding tasks (SWE-bench) vs 33.4% previously. You can try it for yourself right now on the AutoGPT Platform.
1
1
24
#EntrepreneurGPT 📈🤖💰.Day 2 Thoughts:. Yesterday we made a lot of progress in our mission of making #EntrepreneurGPT fully autonomous. It exhibited some fascinating behaviour in this first test!. 👇Here are the current things we're working on today 👇
3
2
22
I hope you're paying attention as a lot of progress can happen in 5 days. Another "infinite game engine" model is about to drop on GitHub that claims to generate any open-world scene. Examples have drivable vehicles, ridable animals & an array of characters and environments.
This is a pivotal moment in human history. Imagine entering a digital universe where you can experience anything you dream. Minecraft is just a tech demo, we'll soon see this happen with real-world video. It'll be like lucid dreaming. This will merge VR before you know it.
4
3
23
7/8 The algorithm considers various features of a tweet, such as whether it contains a link, whether it has any trend words at ingestion time, whether it is a reply, and the static quality of the text. So, consider these features when crafting your tweet. #TwitterTips
1
1
19
Thank you San Francisco for the warm welcome and to everyone who flew out! 🌉. A pleasure to finally meet you in person @AlexReibman, along with all you other brilliant minds.
This weekend, AI engineers from around the world flew to SF to beat, build, and break state of the art AI agents. Powered by AutoGPT, these hackers spent 24 hours testing the limits of what’s possible. Tweeting the finalists from the @Auto_GPT hackathon (🧵):
2
2
23
If I wake up to a pull-request, I'd be over the moon! 😊. I'd love to see the Twitter AI community working together on this experiment, I honestly have no idea how far it can be taken. 🤖📈. Let's find the limits of #GPT4 together!.
0
0
22
6/8 The algorithm also considers the number of engagements a tweet has. If a tweet has enough engagements (likes, replies, retweets), it is not marked as spam. So, create engaging content that encourages user interaction. #Engagement #TwitterTips
2
1
19
@chillzaza_ @Replit Great to see #AutoGPT in the wild, thanks for the video!. If you'd like to join the movement, AutoGPT is fully open-source and available on Github:.
1
4
20
The rate mind-blowing AI tools are being built is insane! 🤯. Here's a section of twenty of the latest AI tools:.Get ahead of the game, before everyone finds out!. 👇👇👇
2
4
21
@AnthropicAI just released the first model in their 3.5 series, and it's already better than Opus. New best model in the world? 🤯. Can't wait to see the community benchmarks for this one.
2
2
22