Command your Conversation with ChatGPT

PLUS: $100 million AI Supercomputer, Google AI Tool for News, Music from Brain Activity

Hey there 👋

A couple of days back Stanford University and UC Berkeley released a research paper about GPT 3.5 and GPT 4’s performance going down. Yesterday, OpenAI released custom instructions in ChatGPT where you can set your preferences which ChatGPT will remember and respond accordingly. Is this a response by OpenAI to the complaints? Nevertheless, it’s going to help to control the output.

Further, Google has pitched its AI tool for writing news articles to a few publication houses. With the Gizmodo Media fiasco, AI at least right now can not be entirely trusted for putting out error-free information. This just reinforces the idea that AI will not replace jobs but people using AI will!

More exciting updates are ready to be uncovered. Let’s roll!

This issue covers:

• Latest Developments 🌍

• Tools of the Trade ⚒️

• Trending AI Jobs 💼

• Prompt of the Day 📜

• Hot Takes 🔥

• AI Meme of the Day 🤡

Read time: 3 mins

Latest Developments 🌍

• My ChatGPT, My Way: OpenAI has rolled out custom instructions for ChatGPT where users can set what ChatGPT needs to know about them and how they’d like ChatGPT to respond, giving more control over its responses.

• GPT March Models Extended: As per the OpenAI blog, GPT 3.5-turbo-0301 and GPT 4-0314 models released in March were scheduled to be retired by September 13, 2023. However, based on developers’ feedback, the company has extended support for those models until at least June 13, 2024.

• AI and Genuine Journalism: Google is testing an AI tool called "Genesis" to write news articles, pitching it to publications like The New York Times and The Washington Post, to assist journalists with tasks and save time.

• OpenAI Backs Government Licensing: An internal memo reveals that OpenAI supports the idea of government licenses for advanced AI systems, aiming to avoid burdening smaller startups with excessive regulations.

• White House Pushes for AI Safety: Companies including Microsoft, Google, and OpenAI, will pledge to abide by certain safety measures till Congress passes AI regulations. The measures include independent expert testing for safety, investing in cybersecurity, sharing safety information, and watermarking AI-generated content. Read OpenAI’s blog for more details.

• Key Departure at OpenAI: Dave Willner, OpenAI's head of trust and safety, and an industry veteran, has stepped down after 1.5 years in the role, to transition to an advisory position, citing the demands of the job since the launch of ChatGPT.

• World’s First AI Supercomputers: Cerebras Systems signed a $100 million deal with UAE's G42 to provide the first, of what could be up to nine, AI supercomputers called Condor Galaxy, as cloud computing providers are searching for alternatives to Nvidia’s chips which are short in supply.

A view of Condor Galaxy

• Context-aware Coding Assistant: GitHub has released GitHub Copilot Chat beta for organizations, a context-aware conversational assistant for developers in the IDE.

• Customizing Diffusion Models Got Easier: FABRIC is a training-free approach that leverages iterative human feedback and self-attention layers to customize diffusion models.
  No need to find the perfect prompt, simply use 👍/👎 to let the model know your preferences.

• AI in Real Estate: Proptech startup Termsheet created a generative AI tool named Ethan to help real estate firms analyze property and market data, and draft memos with recommendations.

• AI Honeytrap for War Secrets: Ukrainian women are using AI-generated profile pictures on dating platforms to deceive Russian soldiers, gaining their trust and extracting military secrets to pass on to the Ukrainian military.

• Current State of LLMs: Researchers have documented the open set of problems relating to LLMs and the areas where they have been successfully implemented, to enable ML researchers to apprehend the current state of LLMs.

• US Military Falling Behind China in AI: During a House Armed Services Subcommittee hearing on AI, Alexandr Wang from Scale AI pointed out that China is outpacing the U.S. in AI development by spending three times more on AI tools, viewing it as their "Apollo project" with the potential to disrupt warfare.

• Music Reconstruction from Brain Activity: Google’s Brain2Music reconstructs music from human brain activity using fMRI and MusicLM, generating music that resembles the musical stimuli experienced by human subjects in terms of genre, instrumentation, and mood.

---

Tools of the Trade ⚒️

• instaVerse by ilumine AI: Set depth map and convert any 2D image to 3D.

• Arvin: GPT-4 assistant wherever you want, for writing, summarising videos and web articles, translation, chatbot, and more.

• Belva: AI outbound phone calls to anyone for any objective, with just 5 lines of code.

• Exspanse: AI developments platform for fast and efficient custom AI model development and deployment, with powerful GPUs and AI code assistance

• Ubie: Tell your symptoms with a 3-min questionnaire and AI will generate a report on possible causes and treatment information.

• Soon: Automate crypto investment with dollar-cost averaging to buy and sell crypto, transfers from/to bank, set a standard portfolio, set a portion aside for tax.

• Flexberry: AI assistant for Business Analysts and UI/UX designers for text analysis, and generating mockup forms, SQL scripts, and UML diagrams.

😍 Enjoying so far, TWEET NOW to share with your friends!

---

Trending AI Jobs 💼

• Amazon: Sr. Applied Scientist | Palo Alto, CA (Apply Now)

• Salesforce: Staff Data Scientist | Dallas, Texas (Apply Now)

• Discord: Software Engineer, ML | Remote, USA (Apply Now)

---

Prompt of the Day 📜

AI as a Dietitian

As a dietitian, devise a personalized meal plan for a [age]-year-old [gender] weighing [weight in kilograms], and a height of [height in centimeters]. The individual is moderately active, prefers a [dietary preference] diet, is allergic to [specific allergens], and wants to [specific dietary goal]. They have been diagnosed with [specific medical condition] and have these dietary restrictions [list of food restrictions, if any].

---

Hot Takes 🔥

• Dot AI is the new dot com ~Bojan Tunguz

• Intelligence is an emergent property of physics. ~Sam Altman

---

Meme of the Day 🤡

Source

That’s all for today!

See you on Monday with more such AI-filled content. Don’t forget to subscribe and give your feedback below 👇

Real-time AI Updates 🚨

⚡️ Follow me on Twitter @Saboo_Shubham for lightning-fast AI updates and never miss what’s trending!!

PS: I curate this AI newsletter everyday for FREE, your support is what keeps me going. If you find value in what you read, share it with your friends by clicking the share button below!