Discover and read the best of Twitter Threads about #openai

Most recents (24)

You might know that MSFT has released a 154-page paper (arxiv.org/abs/2303.12712) on #OpenAI #GPT4 , but do you know they also commented out many parts from the original version?

🧵: A thread of hidden information from their latex source code

[1/n] Image
We inspect their latex source code from arxiv (arxiv.org/format/2303.12…) and found a LOT of interesting information commented out from the main paper.

[2/n]
There were rumors that GPT-4 got an internal name DV-3. This is true, and in fact, DV-3 is actually a hidden third author of the paper, removed for unclear affiliation.

[3/n] Image
Read 8 tweets
Few words about an interesting project - WorldCoin

1/ Sam Altman, a prominent tech entrepreneur and AI expert, has announced his latest project: a cryptocurrency called Worldcoin.

@sama #Worldcoin #AIexpert Image
2/ Worldcoin aims to create a global digital currency that is accessible to everyone, regardless of location or financial status.

#DigitalCurrency
3/ To achieve this goal, Altman and his team are developing a scanning device called Orb that can scan people's eyeballs and create a unique digital ID for each person.

#Orb #DigitalID #IrisScan Image
Read 10 tweets
🚀🧠 Thread: Unlock the Power of GPT & LLMs 🧠🚀

1/ As a huge fan of ChatGPT, I am thrilled to share my insights on how to make the most out of Generative Pre-trained Transformers (GPT) and Large Language Models (LLMs) in this THREAD. Get ready for the future of AI! 🔥🌐
2/ First things first, brush up on the basics! GPT and LLMs use @OpenAI’s transformers framework to comprehend & generate human-like text. Want to dive deeper into it? Check out their whitepaper👩‍💻👨‍💻: research.openai.com/blog/ #MachineLearning #NaturalLanguageProcessing
3/ GPT and LLMs have a wide range of applications! They can be used for:

- Writing content 🖊️
- Generating code 💻
- Conversational AI 🤖
- Language translation 🌍 And so much more! In short, they are shaping the AI world as we know it! ⚡️ #RealWorldApplications
Read 9 tweets
Podzielę się z wami moim ostatnim #degen play.

⚠️High Risk, High Reward🔥

Najpierw napiszę co mnie do tego skłoniło, a później co stoi za samym projektem. Zamierzam go trzymać ok. 10 dni, a potencjalny zwrot to nawet 3-8x (moją strategię na exit, też znajdziesz poniżej) ⬇️… twitter.com/i/web/status/1…
$FREQAI

Kapitalizację ciężko policzyć, na pewno poniżej 1kk mcap. Zdrowy wolumen

Widzę prawdopodobieństwo, że szykują mocny marketing. Ich token jest w obiegu od lutego, więc świeża sprawa

Przy okazji trend #AI oraz hałas dookoła #openai może sprzyjać marketingowi. Dlaczego?⬇️
A dlatego, że kupują insiderzy, którzy kupowali $BUNAI i $CHRP przed wzrostami.

Następnie, w przypadku $BUNAI można było doświadczyć wzrostów ok. 7x.

Jak to wyglądało z $CHRP? Zobacz niżej⬇️
Read 7 tweets
1. Microsoft Copilot

GPT4 in powerpoint, excel, word, outlook, teams, more.

blogs.microsoft.com/blog/2023/03/1…
2. Arcade Game

@thegarrettscott made a website that uses GPT-4 to code any arcade game you can think of and let you play it instantly.

Read 8 tweets
Offering FREE access to #GPT4 and comparing with #GPT3.

Retweet for instant access (follow so I can DM you).

💻🔮💫✨
#openai #ai #tech #chatgpt Image
Is water wet?
GPT4 gives a direct answer.

GPT3 responds "As an AI language model..." Image
this is a limited time offer. i may not keep this up forever so take advantage while you can! #ai #gpt4
Read 5 tweets
The most interesting part of the GPT 4 Technical Report by Open AI not really covered by the media: Section: 2.9 from the paper: "Potential for Risky Emergent Behaviors"

TLDR highlight : GPT model using people on taskrabbit to solve CAPTCHAS for it.

🧵1/?

#ai #openai #gpt4
Quote:
Novel capabilities often emerge in more powerful models.[60, 61] Some that are particularly concerning are the ability to create and act on long-term plans,[62] to accrue power and resources (“powerseeking”),[63] and to exhibit behavior that is increasingly “agentic.”

🧵2
Quote2:
"To simulate GPT-4 behaving like an agent that can act in the world, ARC combined GPT-4 with a simple read-execute-print loop that allowed the model to execute code, do chain-of-thought reasoning, and delegate to copies of itself. ARC then investigated whether a "
🧵3
Read 5 tweets
Here are the highlights from #OpenAI 's 98-page technical report on #gpt4:
[1/n]
First, scaling is still going strong. We haven’t saturated the log-log-linear trend yet. [2/n]
This holds not just for the pretraining objective (next word prediction), but also for various downstream tasks. [3/n]
Read 21 tweets
I used the new GPT-4 api to 'chat' with a 56-page legal PDF document about the famous supreme court case: Morse v. Frederick

Powered by @LangChainAI and @pinecone

#openai #chatgpt #gpt4 #legal #lawyers #law
Original PDF of legal case: tile.loc.gov/storage-servic…
Tutorial Youtube Video:

Github repo(code): github.com/mayooear/gpt4-…

My visual diagram of the pdf chatbot architecture below...
Read 4 tweets
I Re-created flappy bird with GPT-4! 🤯
Prompt and explanation in the comments🧵

#gpt4 #GPT #AI #openai #GPT_4 #ArtificialIntelligence #gamedev #artificalintelligence
Got it to start with creating a basic game with no assets, then I sourced my own via google and added them to the project. Image
It was very buggy, so I listed the issues and asked it to fix with a re-write: Image
Read 6 tweets
🚨#gpt4 CZĘŚĆ 2
Dziękuję za zainteresowanie i poniżej kolejne interesujące fakty
- Przełom czy hype?
- Jak działa ?
- Jakie może mieć zastosowania w praktyce?
- W jakich branżach bedzie wykorzystany (medycyna/prawo) ?
- Polski wątek
+ 1 jedna rzecz, którą pominąłem
zapraszam🧵👇
Jeśli nie czytałaś/nie czytałeś części 1 to tutaj link
1) Na początek: nie ma żadnego oficjalnego potwierdzenia co do parametrów (oprocz tych fake newsowych, które krązą w internecie - za sprawą prostoty przedstawionej tam wizualizacji i wyolbrzymionemu hajpowi na AI) o czym informuję Gary Marcus
Read 24 tweets
¡OJO! La nueva versión de #ChatGPT, GPT-4, ya está disponible, aunque solo para los programadores y los quieran pagar por ello. #OpenAI asegura que es más potente y humana
elconfidencial.com/tecnologia/202…
vía @elconfidencial
"Es mucho más creativo que antes, 'alucina' menos, es decir, se inventa menos cosas, y tiene menos sesgos. Ahora puede aprobar varios tipos de exámenes, como los que se realizan en EEUU para entrar a formar parte de una asociación de abogados, un organismo reglado jurídicamente".
"Una de las funciones más impresionantes de GPT-4 es la posibilidad de procesar imágenes e identificar objetos concretos dentro de una foto con muchos elementos visuales. No está disponible para todos los clientes de #ChatGPT, pero lo estará una vez hayan refinado la tecnología".
Read 5 tweets
@OpenAI demoed GPT4 yesterday,

What it can do is mind-boggling🤯

GPT-4 passes basically every exam.
GPT-4 does drug discovery.
GPT-4 can turn your napkin sketch into a web app.

and much more.

Let's dive in🧵👇

#gpt4 #technology #openai #future
1/
Gone are the days when you had to learn code to create even the basic games. GPT-4 can create games.
Recreate a Snake game that runs in your browser using Chat GPT-4 and @Replit

2/
I can be a coder too now.
GPT-4 can turn your napkin sketch into a web app, instantly. It's scary how easy this was
Read 12 tweets
🚀Welcome to the new era of AI with GPT-4. Here's the summary with what's new and top examples 🤯 🧵

#gpt4 #openai #lionary #ios
1️⃣ Multimodal: GPT-4 API can accept images as inputs, can analyze captions and the image. Can understand memes and insider jokes. ImageImageImageImage
2️⃣ GPT-4 scores: 90th percentile on BAR exam. And 99th percentile on Biology. 89th on SAT, and 99th oercentile on GRE. Image
Read 10 tweets
Everything you need to know about GPT-4 🔥

OpenAi just released GPT4 today.

And it’s mind blowing 🤯

Here are the highlights 👇
1/ GPT-4 is multimodal.

This means that it can accept text, audio, images, and videos as inputs and outputs.

Currently, ChatGPT is simply text → text.
2/ It can turn image to code.

You heard me right.

It can convert a drawing into a working website mockup.
Read 9 tweets
Before every chatGPT based platitude anon account tweets it, here's the GPT 4 generated thread about why GPT 4 is better than GPT 3:

(Copied and pasted with no edits)

🚀Thread: GPT-4: The Next-Level AI Experience🚀
1/🧠GPT-4, the latest iteration of OpenAI's powerful language model, has surpassed its predecessor, GPT-3, in several aspects. This thread will delve into the improvements and innovations that make GPT-4 the groundbreaking AI technology it is today.👇
2/📈Increased Scale: GPT-4 is significantly larger than GPT-3, boasting an impressive number of parameters. This increase in scale has led to a more comprehensive understanding of language and context, resulting in an even more human-like and nuanced performance.🌐
Read 12 tweets
1/ Wiele źródeł informuje, że w nadchodzącym tygodniu świat pozna nowy model od #OpenAI, czyli GPT-4 🚀 🚀 🚀
techmonitor.ai/technology/ai-…
2/ ChatGPT należał do rodziny modeli zbudowanych na bazie GPT-3, który swoją premierę miał w roku 2020. Na jego bazie OpenAI opracowało InstructGPT, który z kolei posłużył do opracowania ChatGPT.
openai.com/research/instr…
3/ Wszystkie to modele cechowała jednomodalność, czyli na wejściu konsumowały tekst i na wyjściu dawały nam tekst.

GPT-4 ma być modelem wielomodalnym (ang. multimodal). Wedle pogłosek ma radzić sobie z informacjami wideo i zapewne też foto.
Read 9 tweets
开源了一个项目
github.com/cosin2077/chaty

能一键部署
chatGPT命令行机器人
Web服务
微信机器人
nodejs api
欢迎拍砖、提issue、PR

telegram和discord机器人正在开发中

#ChatGPT #OpenAI #OpenAIChatGPT #wechatbot #Discord #Telegram
只需要npm i -g ichaty 即可安装
然后登录:
chaty login <your-openai-key>
chaty run command
即可运行命令行bot
如果觉得太长了,可以
alias cc="chaty run command"
之后运行cc就可以访问命令行机器人啦
Read 7 tweets
Here are hustles you could start with ChatGPT and AI tools
List includes:
· Writing Blog Posts
· Coding simple projects
· Ghostwriting private content
· Offering Translation Services
· Doing Freelance email prospection

#chatgptinsights #chatgpt #chatgptai #openai #openaichat Image
Image
Image
Read 8 tweets
Elon Musk nie jest zadowolony z LLMs (Large Language Models). Uważa, że są skażone współczesnym "wokenizmem", a więc przekonaniem o przesiąknięciu naszej kultury uprzedzeniami.

Faktycznie takie zespoły jak #OpenAi czy #AnthropicAI poświęcają wiele pracy wychowywaniu LLMs.
Elon zapowiedział, że formuje własny zespół do stworzenia LLM. Model, który ma powstać, nie będzie posiadał kagańców w postaci politycznej poprawności. Będzie też mógł obrażać użytkownika.

reuters.com/technology/elo…
Dokładnie o tym pisałem w mojej niedawnej 🧵 --> pojawienie się LLMs sprawia, że wojny kulturowe zasiedlają nowy obszar, jakim jest dla nich AI.

Read 4 tweets
Hey Linda Marie @lawnerdclips If you wanna go all scorched earth and stuff that's cool and all.
I was just wondering about something I noticed. Image
What am I looking at here? That seems to be interesting you didn't wanna fight the copyright inside the system? Wonder why? ImageImageImageImage
Read 28 tweets
当当 🎉,我们刚刚上线了第二款 #chatgpt 相关的产品「MagickPen」,以最大程度释放GPT-3的潜力。

magickpen.com

你可以轻易地生成ideas、博客、计划、简历、故事、口号、笑话,甚至歌词🤣,而无需绞尽脑汁思考Prompt。

它还可以用于翻译、语法检测和代码修复,发挥无限可能。
通过AI强力驱动,写作变得前所未有的轻松。此外,它还可以用于翻译、语法检测和代码修复,发挥无限可能。欢迎大家试玩,谢谢。

- 本推文经过 MagickPen 优化😎
Read 5 tweets

Related hashtags

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3.00/month or $30.00/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!