- Superhuman AI
- Posts
- The $140 BILLION AI Gamble
Read time: 2.5 minutes
Welcome back, superhuman
It’s a day of big bets and big bluffs in AI as SoftBank wows to double down on AI despite having little to show for the $140 billion (yes, billion with a b) that they’ve invested in AI.
Grab your knives and forks, we’re serving a treat today.
TODAY’S MENU
Insight: Will the Son rise again after SoftBank’s $140B miss?
News: OpenAI creates a new team to tackle ‘superintelligent’ AI systems
5 AI tools to supercharge your productivity
Research Roundup: Transformative AI, AI infiltration, and LLMs
AI Tutorial: How to make your photos weirder
AI Images of the Day: Coca-Cola cans in different countries
INSIGHTS
Will the Son rise again after SoftBank’s $140B miss?
Image source: Gizmochina
SoftBank CEO Masayoshi Son has a 140 billion problems and not investing in a company at the forefront of the AI boom is presumably every single one of them.
Despite investing an eye-watering $140 billion across 400+ AI startups, SoftBank can only count 1 of the 26 generative AI startups valued at more than $1 billion in their portfolio, according to PitchBook.
But Son isn’t hanging up his boots anytime soon. The company is strategizing for future investments, indicating a continued belief in the potential of AI.
The story of SoftBank's AI investment journey serves as a reminder of the high-risk, high-reward nature of investing in emerging technologies. And the company's willingness to try again highlights its track record as one of the most aggressive investors in tech.
🗞️ Other insights
🧑💻 AI helps in retaining (not replacing) developers (requires Forbes subscription)
TOGETHER WITH PETAL
ChatGPT for document analysis — trusted by MIT libraries
Upload any text document and get answers quickly by talking to your personalized ChatGPT. Chat with one or more documents, extract information from them en masse, summarize, compare, and analyze them with generative AI. Collaborate with your team and share files, annotations, and comments. ISO 27001 certified.
See how it works below:
If can’t access the video above, click here
NEWS
OpenAI creates a new team to tackle ‘superintelligent’ AI systems
OpenAI has formed a new team — the Superalignment Team — consisting of scientists, engineers, and researchers to manage the risks of a potential super-intelligent AI system that could exceeds human capabilities.
OpenAI believes that we're only a decade or less away from the creation of an AI that outsmarts humans, and we need to start managing the risks of such a future scenario right now.
OpenAI says they’re dedicating 20% of the compute they’ve secured to date to this effort
🗞️ Other key events
TOOLS
5 AI Tools to Supercharge Your Productivity
Webcam Effects: A Chrome extension to add, replace, or blur out the background of your video during calls.
QoQo: A Figma plugin to help UX designers create impactful user experiences.
Upside Jobs (sponsored): Finding the right engineer is tough, but not with Upside Jobs. Click here to interview a top ML engineer or AI researcher by the end of the week.
Soundbite: Create human connections with TikTok-style internal communications.
Taleblocks: Transform your text into short, highly-shareable branded videos automatically.
RESEARCH ROUNDUP
The most interesting AI research we found this week
Why transformative artificial intelligence is really, really hard to achieve
This article explains why achieving transformative artificial intelligence (AI) is challenging. It argues that while AI has the potential to revolutionize many areas, there are still significant technical, social, and economic hurdles to overcome. AI needs to transform all essential economic sectors and steps of the innovation process, not just some of them. The article also highlighted that despite progress in some AI subfields, major technical challenges remain, such as fine motor control and learning a causal model. Read more →
Personality traits in Large Language Models
Researchers have been playing psychologist with AI, specifically Large Language Models (LLMs). They've developed a way to measure, analyze, and even shape the 'personality' these AIs show in their text outputs. The kicker? It's not just theoretical - they've found some LLMs can reliably mimic specific personality profiles. It's a game-changer with a whole host of potential applications, but it's not without its ethical head-scratchers. Read the paper →
Other notable research insights in AI
AI TUTORIAL 💡
How to make your photos weirder with Midjourney
Want to make your photos weird but not too weird?
Try this prompt in Midjourney:
closeup of a girl with intricate graffiti snake facial tattoos, in the style of pop culture mashup, neon colours , photorealistic, dynamic lighting --s 1000 --c 15 --weird 0Here’s the result:
-- weird 0
Now using the same prompt but with setting weird parameter to --weird 3000
closeup of girl with intricate graffiti snake facial tattoos, in the style of pop culture mashup, neon colours , photorealistic, dynamic lighting --s 1000 --c 15 --weird 3000
-- weird 3000
Tool used: Midjourney
📈 Feature your product in the world’s biggest AI newsletter
Superhuman is the world’s biggest and fastest-growing AI newsletter with 325,000+ readers working at companies like Apple, Meta, Amazon, Google, Microsoft, and many more. July ad spots are almost sold out. You can book your ad spot by clicking here or below.
🧞 Your wish is my command
What did you think of today's email?Your feedback helps me create better emails for you! |
Thanks for reading.
Until next time!
p.s. if you want to sign up for the Superhuman newsletter or share it with a friend, you can find us here