It’s been a super packed week, AI-wise and otherwise. Donald Trump got shot at and survived, and as an aftermath, Elon Musk pledged his support to the presidential candidate and will also reportedly give $45 million a month to the pro-Trump Super Pac. How is this relevant to today’s newsletter?

Musk, kind of a central figure in the development of AI in the US, could potentially have a say in regulations if Trump becomes president. Republicans have already laid out their plan to scrap Biden's executive order on AI, which has worried some AI experts.

There are also some concerns brewing over the last few months, along the lines of is-the-AI-bubble-going-to-burst-soon? Worries are streaming in from big capital venture firms like Sequoia Capital, which has expressed concerns about the sustainability of the AI industry. 

Last month, David Cahn, partner at Sequoia Capital, stated that the AI sector needs to generate $600 billion in annual revenue to remain viable, a significant increase from his $200 billion estimate in September. 

It’s worth noting that Sequoia Capital’s investments in AI startupsrose from about 16% in 2022 to 60% in 2023. Goldman Sachs has also raised skepticism about generative AI, publishing a report titled "Gen AI: Too Much Spend, Too Little Benefit?" last month.

Will AI meet the same fate as crypto? It’s true that the fruits of the billions of dollars of investment in the technology will be reaped in years or even decades to come. Like crypto, AI has garnered significant attention and investment. I mean, Nvidia became the world’s most valued company last month, although it’s now slipped to the third position again.

I can’t help but compare the trajectories of the two. The rapid rise in cryptocurrency prices was driven by speculative investment, with many investors buying in without fully understanding the technology or its potential applications. This led to inflated prices that eventually corrected.

It’s kind of the same with AI, isn’t it? There is the looming promise of artificial general intelligence (AGI)—human-level intelligence—that the companies are working towards and investing more and more money into.

In fact, according to a recent Bloomberg report, OpenAI told its employees recently that the company believes its technology is nearing AGI and has a five-step classification system to track its AGI progress. OpenAI believes it’s currently at the end of Level 1.

The system ranges from basic conversational AI (Level 1) to advanced reasoning systems (Level 5), capable of performing complex tasks similar to those of a highly educated human without external tools. OpenAI anticipates reaching Level 2, termed "Reasoners," soon.

Another week, another instance of OpenAI finding itself in muddy waters. According to a Washington Postreport, OpenAI whistleblowers have filed a complaint with the Securities and Exchange Commission (SEC), alleging that the company unlawfully restricted employees from reporting potential risks of its AI technology to regulators. 

The complaint alleges that OpenAI enforced overly strict employment, severance, and nondisclosure agreements. These agreements reportedly barred employees from reporting concerns to federal authorities without company consent and waived their rights to whistleblower compensation.

And those were the significant events from last week. Now, moving on to a couple of other AI-related announcements:

Amazon eyes nuclear tech to fulfill AI's energy appetite 

This comes as the global web-shipping company smashed its renewable energy goal seven years ahead of schedule, it said in a report. AWS also purchased a nuclear-powered hyperscale data center campus in March.

Since setting its sights on covering 100% of its global operations, including data centers, with renewable energy by 2030, Amazon has become the top corporate buyer of wind and solar power worldwide since 2019.

Poe lets users create and share web apps

Poe, Quora's subscription-based AI chatbot aggregator, has launched Previews. Previews allows users to create interactive apps directly within chats with AI chatbots like Anthropic's Claude and OpenAI's GPT-4o. These apps can include data visualizations and games, integrate multiple chatbots, and utilize information from uploaded files, including videos. 

It is similar to Anthropic's Artifacts, which is limited to its own models. Previews supports HTML output with CSS and JavaScript functionality.

Teens suspended after creating naked AI images of classmates

In a case highlighting the abusive potential of deepfake and AI technology, 15 teenagers got suspended for a year by a Spanish youth court for spreading AI-generated nude images of their female classmates in WhatsApp groups.

They were charged with creating child sex abuse images and moral integrity offenses. They must also attend classes on gender equality and responsible tech use. The incident caused significant distress among the victims for several months as the photos were circulated on WhatsApp.

‘Hollywood-grade’ text-to-video model: Odyssey AI

Odyssey, the newest AI video tool on the scene, claims to be ‘Hollywood-grade visual AI,’ similar to OpenAI's upcoming Sora tool. 

Co-founder Oliver Cameron, on X, highlighted that users can edit and control the video output. The company is collaborating with multiple large language models to generate distinct layers of the video, allowing separate control over each layer. 

The startup raised $9 million in seed funding led by Google Ventures to develop the video editing tool.

OpenAI is reportedly working on a ‘deep research’ project called ‘Strawberry’

OpenAI is working on a project called Strawberry, which would offer models capable of delivering advanced reasoning capabilities. Does anyone remember project ‘Q’, which Reuters reported last year as a big breakthrough in OpenAI’s goal towards AGI?

Yeah, Reuters now says the same project has been rebranded as ‘Strawberry’. OpenAI hasn’t commented on the latest development, as it remained tight-lipped when project ‘Q’ came to the fore.

Did a friend forward this e-mail to you?

IE₊ SUPPORT INTERESTING ENGINEERING
_{Invest In Science And Engineering}

Enjoy exclusive access to the forefront of AI content, highlighting trends and news that shape the future. Join a community passionate about AI, delve into the latest AI breakthroughs, and be informed with our AI-focused weekly premium newsletters. With IE+, AI reporting goes beyond the ordinary - and it is Ad-Free.

NEWS

• OpenAI agreements blocked staff from revealing AI risks: Whistleblowers
  The agreement discouraged staffers and investors from expressing to federal authorities regarding safety violations.
  

• Japan: Loverse dating app offers ‘Her’ like AI dating bots for loners
  To help the loneliness crisis in Japan, Loverse is a new AI-dating app that hooks up real users with a digital personality.

• Breakthrough AI model uses 93% less time to present perfect battery cells
  According to the company, the new AI solution cuts battery cell design time from two weeks to just a day.

• This man's family will be able to 'engage' with him even after his death
  A platform called ‘Eternos.life’ helps Michael Bommer, who is terminally ill, create a digital self to leave behind after he passes.

MUST-READ

World’s first AI Lego robot head that talks like human gets new life

When Sten decided to build Dave, he didn’t imagine people around the world embracing the world’s first AI lego robot the way they did.

The 17-year-old Dutch teen and the brains behind the Creative Mindstorms put out a video late last year, showcasing Dave’s capabilities.

A marvel made of Lego blocks and gears, Dave boasts as uncanny a resemblance as can be to humans, complete with a wig that makes up for a head full of hair. A pair of eyes that roll at your worst jokes, supplemented by eyebrows that raise adds to the realism.

AI PICTURE OF THE WEEK

We have the world’s first Miss AI.

Kenza Leyli has made history by winning the AI beauty pageant. The AI-powered influencer, with over 200,000 followers on Instagram, was picked from a pool of 1,500 other virtual models in this beauty showdown.

The judges looked at three main things: how realistic the models looked, how advanced the technology was, and how popular they were on social media. They used a points system to give each contestant a final score based on these categories.

There have been some concerns raised that AI might take away jobs and hurt artists, and some think this AI beauty pageant is a strange idea.

So, what’s next? Are we looking at Miss AI Intercontinental? Miss AI Globe? The list is endless, and it seems we’re just at the beginning of turning everything—even beauty pageants—into an AI spectacle.

AI TOOL OF THE WEEK

Yesterday was AI Appreciation Day. Although the day has dubious origins, it serves as a day to evaluate the generative AI revolution's strengths and weaknesses.

For this section, we’re going to discuss the strength of one particular tool that can be very useful for content creators or even media organizations looking to create social media reels out of lengthy videos or interviews. 

Opus Clip leverages AI to automatically generate short, viral-worthy clips from your longer videos. Now, we all know that due to the plethora of content on the internet, humans collectively have the attention span of a mosquito. Opus Clip claims to make capturing and holding an audience's fleeting attention easier.

Here's a step-by-step guide to using it:

1. Uploading your video

• Head to the Opus Clip website (it currently doesn't have a mobile app).

• Click the "Create Clips" button.

• You'll be prompted to upload your video file. Opus Clip accepts various formats like MP4, MOV, and WMV.

2. Customizing your clips (optional)

Before hitting "Get Clips," you can fine-tune some settings:

• Clip Length: Choose the desired length for your short clips. Options range from bite-sized snacks (15 seconds) to longer cuts (3 minutes).

• Target Platform: If you're aiming for a specific platform like TikTok, you can set the clip length accordingly.

• Keywords: Enter keywords you want the AI to focus on. Clips will be generated around those keywords,maximizing their relevance.

3. Opus Clip does its magic

Once you're happy with the settings, click "Get Clips." Opus Clip's AI will analyze your video, identifying engaging moments and stitching them together into short, compelling clips. This process might take a few minutes.

4. Reviewing and editing your clips

• Opus Clip will present you with the generated clips. Each clip will have a title, a short preview, and a "Virality Score" indicating its potential engagement.

• You can play each clip to see if it captures the essence you want.

• If needed, you can edit each clip. Opus Clip allows you to adjust the start and end points for more precise control over the content.

5. Downloading and sharing your clips

Once you're satisfied with your clips, you can download them individually or all at once. Opus Clip provides them in MP4 format, ready to share on your favorite social media platforms.

Bonus Tips

• Use high-quality videos for optimal results.

• Experiment with different clip lengths and keywords to see what resonates best with your audience.

• Opus Clip offers an AI caption generator and emoji suggestion tool to enhance your short clips further.

• While Opus Clip does the heavy lifting, you can still add your own creative touch with editing software if desired.

By following these steps, you can leverage Opus Clip to transform your long-form content into engaging, bite-sized pieces perfect for today's short-form video landscape.