• Forward Future Daily
  • Posts
  • šŸ§‘ā€šŸš€ Microsoft Readies GPT-4.5, Mysterious AI-Designed Chips & Trump Plans Layoffs

šŸ§‘ā€šŸš€ Microsoft Readies GPT-4.5, Mysterious AI-Designed Chips & Trump Plans Layoffs

MIT Tackles AI Ethics, AI-Designed Chips Beat Humans, Microsoftā€™s GPT-4.5 Plans, Robots at Home, Trumpā€™s Layoff Plans, and AI Hiring Surges

Good morning, itā€™s Friday. Microsoft is gearing up for GPT-4.5 and GPT-5, Figureā€™s humanoid robot is learning to assist at home (finally, a roommate who might actually do the dishes), and Trump is planning more layoffs.

Plus, AI and robotics are filling labor gapsā€”but will they replace blue-collar jobs? In todayā€™s Forward Future Original, we explore automationā€™s growing role in manufacturing, elder care, and beyond.

Read on!

šŸ¤” FRIDAY FACTS

Can an AI system forget things? Unlike humans, AI doesnā€™t have memory in the traditional senseā€”but does that mean it never forgets? The answer might surprise you!

Stick around to find out more! šŸ‘‡

šŸ—žļø YOUR DAILY ROLLUP

Top Stories of the Day

OpenAIā€™s upcoming GPT-4.5

šŸš€ Microsoft Preps for GPT-4.5 & GPT-5
Microsoft is gearing up to host OpenAIā€™s upcoming GPT-4.5, codenamed ā€œOrion,ā€ which could launch next week, offering a major upgrade over GPT-4. Meanwhile, GPT-5 is expected by late May, integrating OpenAIā€™s o3 reasoning model for even greater advancements. OpenAIā€™s goal is to unify its AI models, reducing confusion and pushing closer to AGI, making its technology more seamless and powerful across applications.

šŸ¤– Figureā€™s Robot Learns to Assist at Home
Figure Robotics has unveiled Helix, a Vision-Language-Action (VLA) model that allows humanoid robots to understand voice commands and complete household tasks. Helix enables robots to visually assess their surroundings and respond, such as retrieving objects. Unlike traditional robots, it generalizes tasks for unseen items. Still in development, Figureā€™s push into home robotics signals a long-term vision for AI-driven household automation.

šŸ“‰ Trump Plans AI & Chip Agency Layoffs
The Trump administration is moving to cut up to 500 employees at NIST, the agency overseeing AI and semiconductor policy. Commerce Secretary Howard Lutnick is reviewing Biden-era tech programs, including $39 billion in semiconductor incentives and $11 billion in R&D. Part of Trumpā€™s effort to shrink government, the layoffs spark concerns over research setbacks and brain drain. Lawsuits alleging executive overreach have already been filed, challenging the decision.

āš” Global Power Demand Soars Beyond AI
Electricity demand jumped 4.3% in 2024 and is projected to rise nearly 4% annually through 2027, driven mainly by China, India, and Southeast Asia. While AI data centers fuel demand spikes in the U.S. and Europe, the biggest contributors globally are air conditioning, EVs, and heavy industry. Solar and renewables are growing fast, but fossil fuel plants remain dominant. Cutting emissions will require replacingā€”not just supplementingā€”existing power sources.

šŸ’° AI Hiring Startup Mercor Hits $2B
Mercor, an AI-driven hiring startup led by 21-year-old Thiel Fellow Brendan Foody, has skyrocketed to a $2 billion valuation after raising $100 million in a round led by Felicis. Founded in 2023, the company is already profitable, expecting $7 million in revenue and $1 million in profit this month. Investors, eager to secure a stake in AIā€™s rapid expansion, finalized the deal in just two weeks, highlighting the intense demand for AI-driven businesses in todayā€™s market.

ā˜ļø POWERED BY NVIDIA

New GeForce RTX 5090 GPUs Powering AI

nvidia-geforce-rtx-power-adv-ai-badge-rgb

NVIDIA GeForce RTX 50 Series GPUs unlock next level AI capabilities in digital humans, content creation, productivity, and development. From NIM microservices and AI Blueprints, to running local LLMs, image generation and more ā€“ AI is optimized for use and accelerated on NVIDIA GPUs.

Check out the newest technologies NVIDIA unveiled at CES 2025 and watch DeepSeek R1 be put to the test with an RTX 5090.

šŸ“œ ETHICS

Bridging Philosophy and AI: MITā€™s New Ethics of Computing Course

Philosophy and AI

The Recap: MITā€™s new course, Ethics of Computing (6.C40/24.C40), blends philosophy and computer science to tackle the moral challenges of AI and technology. Co-taught by computing professor Armando Solar-Lezama and philosopher Brad Skow, the course challenges students to think critically about ethical dilemmas in AI, internet governance, and decision-making in the digital age.

Highlights:

  • Solar-Lezama invokes the myth of King Midas to highlight unintended consequences in programming, echoing concerns about AIā€™s unpredictable impact

  • Students grapple with AI bias, free will, internet governance, and accountability in autonomous systems, like who is morally responsible when a self-driving car causes harm

  • Professors attend each otherā€™s lectures, adjusting discussions in real-time to ensure interdisciplinary depth

  • A session on the controversial COMPAS algorithm sparks discussion on procedural versus substantive fairness, revealing the complexity of defining bias in AI systems

  • Students like Alek Westover and Caitlin Ogoe explore AI ethics beyond coding, asking deep questions like whether AI should be paid wages.

Forward Future Takeaways:
As AI continues to shape society, courses like Ethics of Computing are crucial in preparing the next generation of tech leaders to think beyond technical feasibility and consider ethical responsibility. By merging philosophical inquiry with engineering, MIT is fostering a new breed of thinkers who can navigate the moral complexities of AI, ensuring that technological advancements align with human values. ā†’ Read the full article here.

šŸ§‘ā€šŸš€ SPREAD THE WORD

Enjoying Fridayā€™s Edition?

Know someone who would enjoy it too? Forward this newsletter or tap the button below to share!

šŸ‘¾ FORWARD FUTURE ORIGINAL

The Blue-Collar Worker and AI

ā€œOne area humans are likely to maintain a relative (or even absolute) advantage for a significant time is the physical world. Thus, I think that the human economy may continue to make sense even a little past the point where we reach ā€œa country of geniuses in a datacenterā€. However, I do think in the long run AI will become so broadly effective and so cheap that this will no longer apply. At that point our current economic setup will no longer make sense, and there will be a need for a broader societal conversation about how the economy should be organized.ā€

Dario Amodei, Machine Loving Grace

The ongoing integration of artificial intelligence and robotics into the labor market is increasingly affecting blue-collar occupations, i.e. jobs that traditionally involve manual labor. While some industries are already seeing significant progress, other areas are still facing challenges that make full automation difficult.

A key issue in this context is the shortage of skilled workers, which poses significant problems for many industries. Demographic change is leading to an aging society, with fewer and fewer young people entering the labor market. At the same time, quality and efficiency requirements are increasing, which in turn increases the demand for qualified workers. In this context, the use of AI and robotics offers a promising solution to close the gap between supply and demand. ā†’ Continue reading here.

šŸ§‘ā€šŸ’» CHIPS

AI-Designed Chips Defy Human Logic but Outperform Traditional Designs

AI-designed chips

The Recap: Researchers at Princeton Engineering and the Indian Institute of Technology used AI to design complex wireless chips in mere hoursā€”an achievement that would take human engineers weeks. The AI-created chips performed better than traditional designs but were structured in ways that human designers couldn't fully comprehend.

Highlights:

  • Instead of using human-engineered templates, the AI employed an inverse design method, working backward from desired performance metrics.

  • The resulting chip structures appeared random and incomprehensible to human engineers but proved more efficient than standard designs.

  • Manufacturing tests confirmed that AI-generated designs outperformed traditional ones in key performance areas.

  • Some AI-generated designs failed, drawing parallels to AI "hallucinations" seen in generative models.

  • The technology could revolutionize not just chip design but broader electronics engineering.

Forward Future Takeaways:
AI-driven chip design is still in its early stages, but its potential to accelerate innovation is undeniable. While human oversight remains necessary to correct errors, the ability to generate highly optimized, novel structures could lead to breakthroughs in computing, telecommunications, and energy efficiency. As AI continues to push past human intuition, the future of hardware design may become increasingly alienā€”yet astonishingly effective. ā†’ Read the full article here.

šŸ›°ļø NEWS

Looking Forward: Stories Shaping the Future

AI coding

šŸ’ø Codeium Raising Funds at $2.85B: AI coding startup Codeium is securing new funding led by Kleiner Perkins, tripling its valuation in six months. With $40M ARR, it focuses on enterprise clients and agentic AI coding tools.

šŸŽ§ Spotify Expands AI Audiobooks with ElevenLabs: Spotify now accepts AI-narrated audiobooks using ElevenLabs, boosting its audiobook library. Authors can generate recordings in 29 languages, though industry debates continue.

šŸ“ˆ OpenAI Hits 400M Weekly Users: OpenAI now serves 400M users weekly, up from 300M in December. Enterprise adoption is booming, with 2M paying business users and developer traffic doubling in six months.

šŸ¤Ÿ NVIDIA Launches AI Tool for ASL: NVIDIA's free platform, Signs, uses AI and 3D avatars to teach American Sign Language with real-time feedback. Initially offering 100 signs, it aims to expand to 1,000.

šŸŖ¦ Europe Risks Becoming AI ā€˜Museum,ā€™ PM Warns: Swedenā€™s leader urges AI deregulation, fearing Europe lags behind the U.S. and China. Leaders push billions in AI investments, but tech firms call for fewer restrictions.

šŸ“½ļø VIDEO

The Industry Reacts to Grok-3 - Speed Is All You Need

Grok 3 debuts as a top AI model, impressing experts with its rapid development and performance. Industry leaders highlight its strong benchmarks, supercomputer speed, and future potential, while early reviews praise its capabilities against top competitors. Get the full scoop in Mattā€™s latest video! šŸ‘‡

šŸ§° TOOLBOX

Effortless Content Creation, Smarter Marketing, and AI-Powered Editing

Luma Dream Machine

šŸŽ„ Luma Dream Machine: Instantly create high-quality AI-generated images and videos with natural motion, seamless looping, and style references.

šŸŽÆ FOMO.ai: Automate your marketing with AI-powered virtual teams that create on-brand content, optimize SEO, and manage social media.

šŸŽ­ AI Face Swapper: Instantly swap faces in videos and GIFs with AI-powered precisionā€”seamlessly preserving expressions, lighting, and movement.

šŸ¤” FRIDAY FACTS

AI Can ā€œForgetā€ ā€“ But Not Like You Think!

While AI doesnā€™t forget in the way humans do, it can lose information through a process called catastrophic forgetting. This happens when a machine learning model is trained on new data, overwriting old knowledge instead of integrating it. Think of it like studying for a new test and accidentally erasing everything you knew about last weekā€™s exam!

To combat this, researchers use techniques like continual learning and memory replay, where old data is periodically revisited to prevent AI from completely wiping out past knowledge. So, while AI doesnā€™t forget because itā€™s distracted or tired (lucky for it), keeping its knowledge intact is still a major challenge!

šŸ—’ļø FEEDBACK

Help Us Get Better

What did you think of today's newsletter?

Login or Subscribe to participate in polls.

CONNECT

Stay in the Know

Follow us on X for quick daily updates and bite-sized content.
Subscribe to our YouTube channel for in-depth technical analysis.

Prefer using an RSS feed? Add Forward Future to your feed here.

Thanks for reading todayā€™s newsletter. See you next time!

The Forward Future Team
šŸ§‘ā€šŸš€ šŸ§‘ā€šŸš€ šŸ§‘ā€šŸš€ šŸ§‘ā€šŸš€ 

Reply

or to participate.