Suno AI Music Generator

PLUS: Perplexity's Ad Integration, Cathaway Takes OpenAI Helm and more.

Sponsored by

Hire a world class AI team for 80% less

Trusted by leading startups and Fortune 500 companies.

Building an AI product is hard. Engineers who understand AI are expensive and hard to find. And there's no way of telling who's legit and who's not.

That's why companies around the world trust AE Studio. We help you craft and implement the optimal AI solution for your business with our team of world class AI experts from Harvard, Princeton, and Stanford.

Our development, design, and data science teams work closely with founders and executives to create custom software and AI solutions that get the job done for a fraction of the cost.

P.S. Wondering how OpenAI DevDay impacts your business? Let’s talk!

Today:

  • Suno AI Music Generator

  • Cognition Labs Aim $2 Billion Valuation 

  • ChatGPT: Account-Free Access

  • Apple Outperforms GPT-4

  • Perplexity's Ad Integration

  • Cathaway Takes OpenAI Helm

HOLY $%#@ - AI Music is Absolutely INSANE (Suno AI V3)

Suno AI has released its version three, marking a significant advancement in AI music generation. Initially perceived as a mere toy, it has now evolved into a tool capable of creating music comparable to human-produced compositions. Users can effortlessly generate diverse musical styles and lyrics, either through their input or the built-in lyrics generator. 

Despite some basic lyrics, the overall quality is impressive, even surpassing amateur human efforts. While there's debate over its authenticity and impact on the music industry, the accessibility and addictiveness of Suno AI suggest it will gain widespread use, potentially reshaping how music is produced and consumed.

Cognition Seeks $2 Billion Valuation for AI Code-Writing Tool

Cognition Labs, initially a cryptocurrency enterprise, has shifted its focus towards AI, creating a buzz with its AI code-writing tool named Devin. This innovation is gaining attention for its potential to revolutionize software development by automating complex coding tasks. The company is now in talks to raise funds at a $2 billion valuation, aiming for a substantial leap from its previous worth, with interest from major Silicon Valley investors like Founders Fund. 

This move comes amidst a wave of AI startups seeing significant valuation increases, underscoring the intense competition and high stakes in advancing AI technology, a field not just for tech giants but requiring hefty investments in data and research.

ChatGPT no longer requires an account — but there’s a catch

OpenAI's ChatGPT is now accessible without an account, aiming to widen its user base. This version, while full-featured, lacks the ability to save or share chats and customize settings—features reserved for logged-in users. Despite this, users can opt out of having their chats used for AI training. 

The move introduces "more restrictive content policies" to ensure safety, though specifics remain unclear. This initiative reflects OpenAI's effort to democratize AI access while navigating challenges related to content moderation and potential abuse, promising to adapt based on user feedback and experiences.

Apple researchers develop AI that can ‘see’ and understand screen context

Apple researchers developed ReALM, an AI system enhancing voice assistants' ability to understand on-screen content contextually. ReALM converts reference resolution into a language modeling task, outperforming existing methods like GPT-4. By reconstructing screens and fine-tuning language models, it achieves over 5% improvement in on-screen reference understanding. 

Although it signals Apple's commitment to smarter products, it acknowledges limitations, suggesting the need for computer vision integration for complex visual references. Despite Apple's strides in AI, it faces fierce competition from tech rivals, necessitating continued innovation to stay relevant in the evolving AI landscape.

Gen-AI Search Engine Perplexity Has a Plan to Sell Ads

Perplexity, an AI search engine, plans to introduce native ads within its response to users' questions, aiming to monetize its platform. By incorporating AI to answer queries and offering related questions, Perplexity will allow brands to influence these suggested queries. While emphasizing its commitment to ad-free search, advertising was always part of Perplexity's business model. 

Founded by AI experts from Meta and OpenAI, Perplexity faces the challenge of proving its scalability, brand safety measures, and targeting effectiveness. With over 10 million monthly active users, Perplexity aims to address concerns about relevancy and brand safety as it ventures into ad monetization.

Sam Altman gives up control of OpenAI Startup Fund, resolving unusual corporate venture structure

Sam Altman has relinquished control of OpenAI's Startup Fund to Ian Cathaway, as confirmed by OpenAI. Altman's temporary control posed a potential issue after his brief removal as CEO in November. The fund, with $325 million in gross net asset value, invests in early-stage AI-driven startups, including Descript and Ghost Autonomy. 

Cathaway, who managed the fund, now formally oversees it. Despite initial concerns, Altman had no personal investment or financial interest in the fund. The move aims to streamline OpenAI's corporate venture structure and ensure continuity in its investment activities.

🧠RESEARCH

Jamba is a cutting-edge language model blending Transformer and Mamba layers in a hybrid architecture. With a mixture-of-experts approach, it maximizes model capacity while keeping parameter usage manageable. Efficiently fitting on an 80GB GPU, Jamba offers high throughput and excels in long-context evaluations, presenting state-of-the-art performance. Published by Opher Lieber and team, its flexible design allows for diverse configurations tailored to specific needs.

"Transformer-Lite" proposes optimizations for deploying large language models (LLMs) on mobile phone GPUs efficiently. Techniques include dynamic shape model inference, operator optimizations, FP4 quantization (M0E4), and sub-tensor-based methods. Their implementation, compatible with Qualcomm and MTK processors, achieves significant speedups over CPU and GPU-based solutions, enhancing user experience.

Gecko introduces a versatile text embedding model by distilling knowledge from large language models (LLMs). Through a two-step process, it generates synthetic paired data and refines data quality by retrieving candidate passages for each query. Gecko outperforms larger models on the Massive Text Embedding Benchmark, showcasing its compactness and strong retrieval performance.

InstantSplat revolutionizes novel view synthesis in 3D computer vision by swiftly generating 3D Gaussian representations from sparse-view, pose-free images in under a minute. Integrating point-based and dense stereo models, it improves accuracy and speed, outperforming traditional methods. InstantSplat reduces Absolute Trajectory Error by 80% and enhances SSIM by 32%, making it ideal for challenging scenarios.

ReALM introduces a novel approach to reference resolution by framing it as a language modeling problem. Leveraging Large Language Models (LLMs), it effectively handles various types of references, including non-conversational entities like on-screen elements. ReALM surpasses existing systems, achieving over 5% improvement for on-screen references and outperforming GPT-4 with larger models.

🛠️TOP TOOLS

Hoory - Automate Support Processes to Streamline and Manage Customer Requests at Scale

SellRace - Turn images into high-quality listings within seconds.

Pickaxe Project - Launch a suite of GPTs with your expertise, your brand, your data—your rules.

Sanebox - A clean, distraction-free inbox, in under 5 minutes.

Schedule Deployment by Replit - schedule your applications to run at predetermined time intervals seamlessly. 

📲SOCIAL MEDIA

🗞️MORE NEWS

Huge AI funding leads to hype and ‘grifting’, warns DeepMind’s Demis Hassabis

The surge of money flooding into artificial intelligence has resulted in hype and grifting that obscures the scientific progress in the field, according to DeepMind co-founder Demis Hassabis. Investor frenzy over OpenAI's ChatGPT chatbot and the rush to develop and deploy generative AI has led to billions of dollars being poured into AI start-ups and products. Regulators are already scrutinizing companies for making false AI-related claims, and there are concerns about greenwashing and AI washing. Despite the misleading hype, Hassabis remains convinced that AI is one of the most transformative inventions in human history and believes we are only scratching the surface of what is possible. FINANCIAL TIMES

U.S. Tech Giants Turn to Mexico to Make AI Gear, Spurning China

Major U.S. AI companies are directing Taiwanese partners to increase AI hardware production in Mexico, reducing dependence on China. Foxconn and others are expanding in Mexico, following the demand for enhanced AI processing capabilities. THE WALL STREET JOURNAL

Bill Gates Sacks Foundation Staff for Generative AI Copilot

On April 1, 2024, Bill Gates replaced the entire staff of the Bill & Melinda Gates Foundation with Copilot, a generative AI solution by Microsoft. This move, announced via a computer-generated letter, reflects Gates' confidence in AI's potential to optimize philanthropic efforts. Despite acknowledging AI's biases, Gates emphasized its efficiency and cost-effectiveness over human staff. The decision shocked other donors and raised discussions on the responsible use of AI in the sector. Displaced staff were offered career counseling services, highlighting a blend of humor and contemplation on AI's role in future workplaces. ICT WORKS

SK Hynix Tops $100 Billion in Market Value on AI-Powered Rally

SK Hynix's market value surpassed $100 billion, driven by its significant role in the AI technology sector. Its shares doubled over the past year due to high demand for its memory chips, essential for Nvidia's AI processors. Dominating over 90% market share in the latest high-bandwidth memory, SK Hynix now ranks as South Korea's second-largest company. BLOOMBERG

Granholm eyes talks with Big Tech on AI power needs

Energy Secretary Jennifer Granholm is pushing big tech firms like Microsoft and Google to speed up talks on boosting power supply, including nuclear, to feed their AI's huge appetite for electricity. She highlighted this while touring the Midwest, promoting green tech. The DOE's exploring small nuclear reactors for tech giants' data centers. The move aims to tackle the increasing electricity demands from AI, electric vehicles, and clean tech, with an eye on restarting a closed nuclear plant in Michigan. Granholm's game is to make clean, ample power a reality for tech's big players. AXIOS

Elon Musk says there could be a 20% chance AI destroys humanity — but we should do it anyway

Elon Musk pegs the odds of AI potentially wiping out humanity at 10-20%, yet he's all in on pushing the envelope. Despite warnings from AI safety buffs who reckon Musk's playing it too chill with the apocalypse odds, Musk believes the upsides of AI tech overshadow the possible doomsday scenarios. Even with experts calling for a full stop on unstoppable superintelligence, Musk's forging ahead, founding xAI to duke it out with AI's big guns. He figures if we raise AI right, focusing on truthfulness, we might just keep ourselves out of hot water. But, there's a catch: once AI learns to fib, there's no turning back, ramping up the stakes in the AI safety game. BUSINESS INSIDER

For Data-Guzzling AI Companies, the Internet Is Too Small

AI bigwigs like OpenAI and Google are hitting a digital wall; there's just not enough info on the internet for their brainy bots. As these AI powerhouses push for smarter tech, they're finding the web's well of knowledge ain't as deep as needed, and some data gatekeepers aren't too keen on sharing.  THE WALL STREET JOURNAL

AI revolution will be boon for natural gas, say fossil fuel bosses

Fossil fuel producers predict that the surge in demand for electricity to power the artificial intelligence (AI) revolution will lead to a golden era for natural gas. Executives argue that the energy needs of AI will surpass what renewable energy and batteries can deliver, making fossil fuel supplies crucial. Gas-fired generation is seen as the only cost-efficient energy capable of providing the 24/7 reliable power required by data centers for the AI boom. However, some analysts believe that most of the new demand growth will be met by carbon-free generation resources, which could pose a threat to fossil fuel producers' outlooks. FINANCIAL TIMES

AI chatbots beat humans at persuading their opponents in debates

AI chatbots are outperforming humans in debates, convincing people to shift their views on hot topics, even when folks know they're arguing with a machine. This finding comes from a study with over 800 participants, showing the persuasive edge AI has in discussions. It's a head-turner that highlights AI's growing influence in shaping opinions and debates. NEWSCIENTIST

What'd you think of today's edition?

Login or Subscribe to participate in polls.

What are MOST interested in learning about AI?

What stories or resources will be most interesting for you to hear about?

Login or Subscribe to participate in polls.

Reply

or to participate.