π€GPT-4 supports image and text input, while GPT-3.5 only accepts text.
π€The GPT-4 has performed comparable to humans in a variety of professional and study tests. For example, it passed the bar exam, placing in the top 10% of test takers.
3/ π€OpenAI spent 6 months testing and configuring GPT-4. In simple chat the difference between GPT-3.5 and GPT-4 is not so noticeable, but on more complex tasks it becomes apparent. GPT-4 is more robust and creative then GPT-3.5, and can handle more complex and intricate request
4/ The green bars on the graph indicate how much better the new model is doing:
5/ π€An active subscription to ChatGPT Plus ($20) will be required to use the new version of GPT-4 now. OpenAI plans to eventually introduce a paid subscription for those who use the system in large volumes, but hopes to leave some free queries for regular users.
6/ Visual input:
GPT-4 now understands not only text but also images: documents with text and photos, diagrams, screenshots and more.
7/ Manageability:
Instead of the classic ChatGPT personalization with a fixed number of words, tone, ChatGPT users will soon be able to customize the style and tasks for their AI by describing these preferences in a "system" message.
8/ Limitations:
The GPT-4 has the same limitations as previous GPT models. It still makes up facts and makes errors in reasoning. The current model outperforms the GPT-3.5 by 40% according to the company's internal estimate.
9/ Risks and mitigation measures:
The team is strengthening the security of GPT-4 through screening and filtering the data before training. Experts were hired to test high-risk queries. Feedback and data from experts in these areas were used to improve the model.
10/ Overall, team interventions have reduced dangerous requests, but there are still situations where users break the algorithm and access dangerous content. The team is bringing in external researchers to assess the potential impact at this stage and in the future.
11/ Predictable scaling:
GPT-4 focuses on creating a scalable deep learning stack. The team has developed infrastructure and optimizations that have predictable behavior across scales. In addition, they have begun to develop a methodology for predicting more interpretable metrics
12/ API:
To access the GPT-4 API (uses the same ChatCompletions as gpt-3.5-turbo) you need to register in the waiting list.openai.com/waitlist/gpt-4β¦
I have a unique opportunity to suggest new game features for @PlayMightyHero from @MightyBearGames game studio, as I am a participant in their #degen100 contest.
1β£Since @PlayMightyHero is a battle royale game, the first thing I would like to propose is a new map idea.
I propose to make a map floating in the air, and instead of the usual tapering circle, use destroying the edges of the map so that latecomers fall down from the map
2β£The second feature will deal with the in-game character. Each character will have its own unique special skill that can be used in battle. For example, one character may have the ability to move quickly, another may have the ability to create temporary barriers, β‘οΈ
How do you analyze projects on Twitter to see if it is a Gem or a Scam?
Here are five tools I use to analyze and find alpha in a Twitter:
2/ @TweetScout_io (tweetscout.io) - is a great tool to find out who is subscribed to whom, and importantly, it determines Twitter accounts' scores based on their internal algorithms.
3/ @TweetDeck (tweetdeck.twitter.com) - allows power users to easily manage their Twitter accounts, customize their tweet feeds, delay posts, and more.