This Week In AI Business #51
At the end of November 2024, we marked two years of the “ChatGPT moment” to make things even more noisy, buzzy, and - if you’re like me - fun, the Big Tech players (now transitioning into the “AI Hyperscalers Club) have speeded up the release schedule for any possible angle.
In short, the AI race, rather than slowing down, as many claimed just a few weeks back, has done the opposite; it’s moving faster!
Yet, the domain on top of which AI players are competing moved from generation to reasoning, with some key moves coming from many of them, highlighting what might be the hot trends coming in 2025!
Let’s go!
The AI Reasoning Race
As “reasoning models” have become the new hot thing in tech, they’ve also pushed the foundational players to move quickly.
Outside the buzz, reasoning is the next frontier of AI, and we’re getting there fast (it will be tricky, but it’s a field that will mature in the coming 3-5 years).
This week, Google released its reasoning model to compete with OpenAI’s o1. In parallel, the ChatGPT maker also launched version two, called “o3” (to prevent trademark infringement claims by the British telecommunications firm O2) of its reasoning model!
Let’s see some of the details!
The OpenAI’s AI reasoning rush!
For one thing, OpenAI is the top player in keeping momentum on its product roadmap.
And that’s impressive considering the massive amount of drama the company went through in the last couple of years, given its meteoric success (with that, drama comes as default).
OpenAI’s 2024 advancements highlight key patterns: expanding multimodal AI with voice, video, and vision capabilities; targeting enterprises through tailored tools and partnerships; addressing ethical and regulatory concerns like privacy and misinformation; innovating models for better reasoning and creativity; and forging global collaborations with media, tech, and educational institutions.
As reported by TechCrunch, OpenAI confirmed the release of o3 and o3-mini reasoning models, claiming advancements toward AGI with step-by-step reasoning, adjustable thinking times, and superior benchmark performance.
While addressing safety with “deliberative alignment,” o3 highlights include a record-breaking 25.2% score on Frontier Math.
The launch follows increased competition and leadership changes within OpenAI.
How did OpenAI roll it out?
• Launch of o3 Models: OpenAI unveiled o3 and o3-mini, successors to the o1 reasoning model, with claims of approaching AGI capabilities in certain conditions.
• Enhanced Reasoning: o3 features a “private chain of thought,” enabling step-by-step reasoning with adjustable thinking time for improved accuracy.
• Improved Performance: Outperforms o1 on multiple benchmarks, including a record-breaking 25.2% on Frontier Math, where no other model exceeds 2%.
• Multifunctional Design: Offers low, medium, and high reasoning time modes to suit varying complexity and response requirements.
• AGI Proximity: Scored 87.5% on ARC-AGI, showcasing progress toward artificial general intelligence.
• Competition: Released amidst a surge in reasoning models by rivals like Google, Alibaba, and DeepSeek.
• Safety Measures: Employs “deliberative alignment” techniques for improved safety and transparency.
• Future Plans: Partnering with ARC-AGI for next-gen benchmarks and staggered release plans starting January 2025.
• Leadership Change: Alec Radford, a key scientist behind OpenAI’s GPT models, departed to pursue independent research.
Google doubles down on reasoning
Google’s experimental Gemini 2.0 Flash Thinking AI model provides step-by-step reasoning for complex problems, integrating visual and textual data for enhanced multimodal understanding.
Competing with OpenAI’s o1, it offers transparency and improved outcomes, leveraging Gemini 2.0’s speed.
Available on Google AI Studio, it advances Google’s push into “agentic” AI solutions.
What are the details there?
• Introduction of Gemini 2.0 Flash Thinking: Google launched a new experimental AI reasoning model designed to break down complex problems into smaller steps, offering detailed explanations of its “thoughts” during problem-solving.
• Enhanced Problem-Solving Capability: Demonstrates multimodal reasoning by integrating visual and textual data for tasks like physics problems and other complex queries.
• Competes with OpenAI’s o1 Model: Positioned as a competitor to OpenAI’s reasoning model, offering transparency and stronger reasoning outcomes.
• Transparency in Reasoning: Provides a step-by-step rundown of its process, improving the clarity and reliability of its answers.
• Built on Gemini 2.0: Leverages the speed and advancements of Google’s upgraded Gemini 2.0 model, enhancing its reasoning performance.
• Accessible on Google AI Studio: Available for experimentation, marking a step forward in Google’s efforts to push “agentic” AI applications.
Google has doubled down on AI search, which might only accelerate its decline!
As I’ve highlighted in the post-Google world, it doesn’t matter how much the company will try; search itself is a UX that won’t work in the current AI paradigm.
And yet, at least to buy time (or maybe accelerate the decline?) Google seems to get ready to doubling down on AI, will this kill search?
As reported by The Information, Google might be planning to implement AI way more aggressively into search via an AI Mode Option, which would turn Google Search into a Gemini AI-like Assistant.
The move might come as competition intensifies and Google’s AI overviews might not be enough to prevent the consumer disruption happening through AI chat interfaces like Perplexity AI and ChatGPT Web Search.
If it confirms that might be the riskiest and boldest move from Google, in 30 years, on its core product.
Why and how?
• AI Mode in Google Search: Google plans to introduce “AI Mode,” offering conversational, chatbot-style answers alongside traditional search results, accessible via a new tab.
• Integration with Gemini-like Chatbot: AI Mode will provide conversational answers with links to external websites and a follow-up question option, similar to Google’s Gemini chatbot.
• Google’s Vision: The AI Mode aims to enhance discovery across the web, leveraging advanced models to make search more intuitive and conversational.
• Growing Competition: Companies like Perplexity AI and Reddit are also incorporating AI-powered conversational interfaces, signaling increased competition in AI-driven search capabilities.
• Perplexity AI Success: Perplexity AI recently raised $500 million, tripling its valuation to $9 billion, with over 100 million weekly queries served.
• Reddit’s AI Experiment: Reddit is testing “Reddit Answers,” an AI conversational interface that curates summaries and links from community discussions.
• OpenAI’s AI Search Features: OpenAI previously integrated prototype AI search features into ChatGPT, aiming to make web searches faster and more conversational.
• Google’s AI Ambitions: Google’s move highlights its commitment to leading the AI-powered search evolution, integrating real-time information with conversational ease.
Will this work? Honestly, I believe it will only further kill a UX, which, at scale, was not thought for the current AI paradigm.
Perplexity joins the big guys club while it steps up its RAG game
Perplexity is stepping up its “RAG game” to enhance AI search.
Indeed, as reported by TechCrunch, Perplexity has acquired Carbon, a startup specializing in retrieval augmented generation (RAG), to enhance its AI search capabilities.
The integration will allow Perplexity’s AI to search through work files in tools like Google Docs and Slack by 2025. This positions Perplexity in the competitive enterprise AI search market alongside OpenAI and Google.
Why does it matter?
• Acquisition Details: Perplexity acquired Seattle-based Carbon, a startup specializing in retrieval augmented generation (RAG), enabling AI systems to access external data sources.
• New Capabilities: Perplexity will integrate Carbon’s technology to enable searching through work files and messages in tools like Google Docs, Notion, and Slack by early 2025.
• Focus on Enterprise Search: This acquisition positions Perplexity to enter the competitive enterprise AI search space, connecting generative AI to large, unstructured corporate datasets.
• Enhanced AI Answers: Carbon’s RAG expertise will allow Perplexity’s AI to provide answers informed by internal databases, cloud storage, and document repositories.
• Competitive Landscape: The move aligns Perplexity with efforts by OpenAI, Google, and others who are developing enterprise search products, inspired by market leader Glean.
• Expansion Strategy: This is Perplexity’s second acquisition, following its 2023 purchase of Spellwise, which contributed to its mobile app development.
In the meantime, Perplexity also (nearly) joined the Decacorn club!
In fact, as reported by Bloomberg Tech, Perplexity AI has officially raised $500M, tripling its valuation to $9B.
Backed by investors like SoftBank, Jeff Bezos, and Nvidia, the AI search startup differentiates itself with real-time capabilities and tools for internal file searches.
It boasts 15M active users, finance-focused features, and revenue-sharing deals with publishers like Time and Fortune.
What drove this massive valuation?
• Valuation Growth: Perplexity AI tripled its valuation to $9 billion after closing a $500 million funding round led by Institutional Venture Partners.
• Investor Backing: Previous backers include SoftBank Vision Fund 2, Amazon founder Jeff Bezos, and Nvidia.
• Competitor Differentiation: Known for offering real-time search capabilities, Perplexity competes with OpenAI, Microsoft, and Google in integrating generative AI into search.
• Product Offerings: Features include tools for searching internal organizational files, finance-related searches (e.g., stock prices), and partnerships with publishers like Time and Fortune.
• User Growth: As of March, Perplexity reported over 15 million active users.
• Revenue Streams: The startup has established revenue-sharing partnerships with publishers, addressing plagiarism concerns.
AI Capex + R&D Is The Hyperscalers Strategic Paradigm
As reported by Axios, a mixture of Capex (to ramp up AI infrastructure) and R&D (AI model integration) has become the staple of the current AI race.
Big Tech (we might well call them “hyperscalers” in this phase) is pouring unprecedented capital into AI data centers, driving growth in energy, semiconductors, and real estate.
Capex spending has surged by tens of billions in 2024, fueling AI infrastructure development.
Leaders like Sundar Pichai prioritize “overinvestment” to future-proof operations, with the potential for repurposing assets and market consolidation ahead.
In fact:
• Tech Capex Surge: Big Tech is spending unprecedented amounts on AI infrastructure, driving growth in real estate, building materials, semiconductors, and energy.
• Energy Demand: Data centers consume power comparable to small cities, boosting the energy sector.
• AI Capex vs. R&D: Capex focuses on data centers and equipment, while R&D targets AI integration and model training.
• Capex Growth: Major tech players have increased capital expenditures by tens of billions in 2024, with continued growth expected in 2025.
• Long-Term Strategy: Capex is viewed as a strategic investment in the future, with potential for repurposing facilities for new technologies.
• Acquisition Phase Two: Initial capex spending may lead to M&A consolidation when the market stabilizes.
• CEO Perspective: Leaders like Sundar Pichai emphasize the risks of underinvestment over overspending in the AI race.
• Versatile Assets: Data centers built for AI can adapt to support future tech trends if the AI bubble bursts.
The AI Cloud Race Will Go Well Into 2025!
A few weeks back, I’ve covered in detail both the new AI Hardware Paradigm:
Why AI Data Centers will be a critical component to enable the developing AI ecosystem, and why we only started building these in the coming 5-10 years:
Well, Microsoft is working way ahead. While other players like Google and Amazon are also building their own vertical AI Chip capabilities, Microsoft is going all in on pushing as much AI CapeX as possible!
Microsoft won the hyperscaler prize for 2024!
As reported by The Financial Times, Microsoft has been by far the largest AI spender this year.
In fact, the company acquired 485,000 Nvidia Hopper chips in 2024, leading AI infrastructure investments with $31 billion in data center spending.
Supporting OpenAI and Azure services, it outpaced rivals like Meta and Google.
While Nvidia dominates the GPU market, competition from AMD and custom AI chips from Google, Meta, and Amazon is rising.
What’s the strategy behind this massive AI spending?
• Microsoft’s GPU Leadership: Microsoft acquired 485,000 Nvidia Hopper chips in 2024, twice as many as rivals like Meta, Amazon, and Google, solidifying its AI infrastructure lead.
• Massive AI Investment: Microsoft has invested $13 billion in OpenAI and expanded Azure infrastructure for AI services like Copilot and external customers.
• Surge in Data Center Spending: Tech companies spent $229 billion on servers in 2024, with Microsoft leading at $31 billion in capital expenditures.
• AI Chip Competition: Microsoft is also developing its Maia AI chips, although it’s earlier in development compared to Nvidia and competitors like Google’s TPUs and Amazon’s Trainium chips.
• Global GPU Market Growth: Nvidia GPUs accounted for 43% of server spending in 2024, while rivals like AMD and custom AI chips from Google, Meta, and Amazon gained traction.
• Rising AI Demand: The increased GPU acquisition supports AI training for models like OpenAI’s o1 and competition against players such as Anthropic, xAI, and Tencent.
• Nvidia Market Challenges: Despite its dominance, Nvidia faces competition from custom chips and geopolitical concerns about US-China restrictions on AI technology.
But in reality, Microsoft’s infrastructure strategy can be summarized as:
Get as many GPUs as you can now, as we don’t know how many we’ll be able to get tomorrow. And let’s make sure Azure (in the AI age) stays in the top three spots among the major cloud providers (Amazon AWS and Google Cloud).
Are you ready for the AI GPU to become like the CPU?
As reported by The WSJ, The AI Cloud market is opening up with native players entering the space.
The latest is Vultr, which raised $333M at a $3.5B valuation, led by AMD and LuminArx Capital, to expand its AI cloud services and GPU capacity.
AMD aims to be Vultr’s preferred hardware provider and support its new AI clusters.
Vultr competes in the growing AI chip market, challenging Nvidia’s dominance while serving major clients.
Why does it matter?
• Funding Round: Vultr raised $333 million in its first outside capital round, led by AMD and LuminArx Capital, reaching a $3.5 billion valuation.
• AI Focus: Vultr plans to expand its AI cloud service, which is expected to become its largest business segment, leasing GPU access for AI workloads.
• AMD Partnership: AMD aims to become Vultr’s preferred AI hardware provider, supplying GPUs like the MI325X and upcoming MI350 for Vultr’s AI clusters.
• Supercomputer Cluster: Vultr is building a large-scale AI cluster with AMD GPUs at its Chicago data center.
• Market Growth: The AI semiconductor market is projected to grow from $117.5 billion in 2024 to $193.3 billion by 2027, with Nvidia dominating 95% of the market.
• Competitors: Vultr competes with Nvidia-backed CoreWeave and other providers like TensorWave, leveraging diverse GPU suppliers, including Nvidia and AMD.
• Customer Base: Vultr serves businesses like Activision Blizzard and Bharti Airtel, supporting both AI and core IT infrastructure needs.
• Broader Industry Trends: AMD and startups are challenging Nvidia’s dominance in AI chips, with cloud companies rapidly pivoting to meet generative AI demands.
As we get into the 2030s, AI Chips will become a commodity (yes, we might not see that before then). And it’s interesting to see how, as 2024 closes, more and more native AI Chip players are coming to the space, willing to take some of NVIDIA’s bonanza!
NVIDIA goes into Small AI Computing
Jensen Huang and the team have been cooking even more than usual!
Not everyone needs a multi-billion AI supercomputer.
Indeed, if at all, only a few massive clients will, mainly the "Hyperscalers" or "AI First Movers" (see Microsoft and other big tech players AI Chip spending spree for 2024, which passed $229 billion!).
All the rest, especially those experimenting with AI/robotics at a small scale, won't!
That’s why NVIDIA’s new Jetson Orin Nano Super Developer Kit offers affordable, compact, generative AI power at $249, delivering a 1.7x performance boost.
It supports robotics, LLMs, and computer vision with extensive software and community support. Accessible to developers, it enables cutting-edge AI innovation, from edge applications to foundational model development.
What does it come with it?
• Affordable Generative AI Supercomputer: NVIDIA introduces the Jetson Orin Nano Super Developer Kit, a compact AI supercomputer priced at $249, significantly down from $499, offering a 1.7x gain in generative AI performance.
• Enhanced Capabilities: Boasting 67 INT8 TOPS performance and 102GB/s memory bandwidth, it supports AI tasks like LLM chatbots, visual agents, and robotics.
• Software Upgrades: Existing Jetson Orin Nano devices can upgrade their software for the same 1.7x performance boost, enhancing accessibility for developers and students.
• Comprehensive AI Ecosystem: Supports NVIDIA’s AI software, including Isaac for robotics, Metropolis for vision AI, and Omniverse for data simulation, reducing development time with pre-trained models and synthetic data generation.
• Developer Support: Extensive tutorials, community projects, and partnerships enhance user experience with tailored tools, sensors, and custom solutions.
• Accessibility for Emerging Developers: The kit serves as an ideal entry point for those exploring generative AI, robotics, or computer vision, aligning with the shift toward foundation models and edge AI innovation.
AI is Benioff’s blessing and curse, as Salesforce is trying to define itself as a disruptor rather than disrupted by AI!
As reported by CNBC, Salesforce plans to hire 2,000 AI-focused salespeople, doubling its previous target, as it ramps up AI product sales.
The next-gen Agentforce AI agent, launching in February 2025, will enhance customer support efficiency.
Salesforce’s AI tools reduce human escalations and position the company as a leader in AI-powered customer service solutions.
How and why is Salesforce ramping up its AI commercial efforts?
• AI-Focused Hiring Surge: Salesforce will hire 2,000 salespeople specializing in AI, doubling its previously announced hiring plans. Over 9,000 applicants have expressed interest in these roles.
• Agentforce 2.0: The second generation of Salesforce’s AI agent technology, capable of answering complex queries in Slack using comprehensive data, will launch in February 2025.
• AI’s Impact on Support: Salesforce’s AI-powered support systems have reduced human escalations from 10,000 to 5,000 weekly, enhancing efficiency.
• AI-Driven Strategy: The company focuses on integrating AI into customer service, sales, and marketing workflows to boost productivity and streamline operations.
• Market Competition: CEO Marc Benioff highlighted Salesforce’s AI advancements while contrasting them with Microsoft’s Copilot tools, positioning Salesforce as a leader in AI-driven customer service solutions.
This is critical, as along with this massive, dedicated AI sales force, the company will be building up its Enterprise AI Agentic Solution, which will be the first use case to tackle AI Customer Support!
And one thing is for sure: we'll see whether Salesforce will be successful at it. But whether Salesforce makes it or not, the whole enterprise ecosystem around AI will develop thanks to these investments!
What’s next for Google (Alphabet) if search will be gone? Robotaxis are coming!
As reported by TechCrunch, Waymo will launch its first international robotaxi testing in Tokyo in early 2025, tackling challenges like left-hand driving and dense urban environments.
Partnering with GO and Nihon Kotsu, the program includes mapping key areas and training drivers. This expansion follows Waymo’s U.S. city tests and contrasts with GM’s Cruise withdrawal.
How?
• First International Expansion: Waymo will test its robotaxis in Tokyo starting early 2025, marking the first deployment outside the U.S.
• Left-Hand Driving Challenges: Tokyo’s unique challenges include left-hand driving and dense urban environments, adding complexity to Waymo’s technology testing.
• Collaboration with Local Partners: Waymo is partnering with taxi-hailing app GO and taxi company Nihon Kotsu, which will manage and service its vehicles.
• Mapping and Training: Nihon Kotsu drivers will manually map key areas in Tokyo while training for Waymo’s self-driving Jaguar I-Pace vehicles.
• Development Program Continuation: This Tokyo deployment builds on Waymo’s “road trips” testing in diverse U.S. cities, adapting to unique conditions like rain in Miami and extreme heat in Death Valley.
• Contrast to Cruise’s Withdrawal: The move follows GM’s decision to end its Cruise robotaxi program, including plans for a similar Tokyo service with Honda.
Out of all the possible parallel universes, where Alphabet is still the most relevant company in the world, Waymo is what might make this parallel universe exist!
In the meantime…
Google enters robotics
The amount of pacing Google has gotten into, considering how much more conservative it has been in the last two years, surprised me!
I’ve touched in detail in Google’s AI Triad what I was impressed about on the latest Google moves!
And that’s not over.
As Google has managed to bring everything under the umbrella of Google’s AI, the fact is, the company still has two steps that walk in parallel and are now finally coming together: The Google Brain Soul and The Google DeepMind Soul!
DeepMind has always been the innovative part of Google, looking at breakthroughs outside the core of the business., In fact, Google has just massively entered the robotics space by partnering with Apptronik.
More precisely, Apptronik, in partnership with Google DeepMind Robotics, will develop AI-powered humanoid robots, combining advanced AI with cutting-edge robotics.
Apptronik’s Apollo robot, designed for industrial tasks, reflects nearly a decade of development.
This collaboration builds on Apptronik’s partnerships with GXO and Mercedes-Benz, aiming to expand applications into healthcare, home assistance, and beyond.
How will the partnership unfold?
• Strategic Partnership: Apptronik teams up with Google DeepMind Robotics to advance AI-powered humanoid robots for dynamic environments.
• Embodied AI Focus: Partnership combines Apptronik’s robotics hardware with DeepMind’s AI expertise to create versatile, intelligent, and safe humanoids.
• Apollo Humanoid Robot: Apptronik’s flagship robot, Apollo, is designed for industrial tasks, standing 5’8” and weighing 160 pounds, with safety and reliability at its core.
• DeepMind’s AI Expertise: DeepMind’s robotics team leverages state-of-the-art AI systems, including Gemini, for reasoning and real-world action.
• Industry Collaborations: Apptronik has partnered with GXO and Mercedes-Benz, with more collaborations expected in 2025.
• Legacy in Robotics: Apptronik, founded in 2016, has extensive experience in human-centered design, having developed 15 robots, including NASA’s Valkyrie.
• Future Applications: The partnership aims to expand humanoid use beyond manufacturing and logistics into healthcare and home assistance.
I’ve already highlighted a few issues, such as how embodied AI is one of the key trends for 2025!
Here is the complete list!
Recap: In This Issue!
AI Reasoning Race:
OpenAI launched the "o3" reasoning models with step-by-step reasoning, AGI proximity (87.5% on ARC-AGI), and superior benchmarks like 25.2% on Frontier Math.
Google introduced Gemini 2.0 Flash Thinking, providing enhanced multimodal reasoning and step-by-step transparency.
Google AI Search Revamp:
Plans an "AI Mode" in search for conversational answers alongside traditional results, leveraging Gemini-like capabilities, competing with OpenAI and Perplexity.
Perplexity AI Advancements:
Acquired Carbon for enhanced enterprise AI search with retrieval-augmented generation (RAG).
Raised $500M, tripling valuation to $9B, with real-time capabilities and publisher partnerships.
Tech Capex Surge:
Big Tech spent unprecedented billions in AI infrastructure (e.g., Microsoft’s $31B, acquiring 485,000 Nvidia chips) to maintain hyperscaler dominance and future-proof data centers.
Nvidia Expands Reach:
Introduced the $249 Jetson Orin Nano Super Developer Kit for affordable, small-scale AI experimentation, supporting robotics and LLMs.
Salesforce AI Hiring Surge:
Doubling AI sales staff to 2,000; launching Agentforce 2.0 in February 2025 to enhance AI-powered customer support systems.
Waymo’s Global Expansion:
First international robotaxi testing in Tokyo, focusing on left-hand driving and dense urban environments.
Google Robotics Push:
Partnership with Apptronik for AI-powered humanoid robots, leveraging DeepMind's advanced AI and Apptronik’s hardware for applications in manufacturing, healthcare, and home assistance.
AI Hardware Market Growth:
Vultr raised $333M for AI cloud services and supercomputer clusters, challenging Nvidia's dominance alongside AMD's increasing market share.
Embodied AI as a Trend:
Both Google and Apptronik reflect growing interest in combining AI with physical robotics, marking a key trend for 2025.
Key Patterns in the AI Landscape Consolidating In 2025
AI Reasoning as a Core Focus:
Shift from generation to reasoning, with major players like OpenAI and Google racing to develop advanced reasoning models.
Emphasis on step-by-step problem-solving, transparency, and enhanced multimodal capabilities (e.g., Gemini 2.0 Flash Thinking, OpenAI’s o3).
Hyperscaler Capex and AI Infrastructure Investment:
Big Tech prioritizes unprecedented capital expenditures (e.g., Microsoft’s $31B in AI infrastructure) to secure dominance in AI capabilities.
Data center expansion, energy consumption, and GPU acquisitions are critical for scaling AI operations.
AI-Driven Enterprise Solutions:
Focus on enterprise-grade AI tools, such as Salesforce’s Agentforce 2.0 and Perplexity’s enterprise search expansion, to integrate generative AI into organizational workflows.
Increasing adoption of retrieval-augmented generation (RAG) for internal and external data connectivity.
Emergence of AI Cloud Ecosystems:
The rise of AI cloud-native players (e.g., Vultr, Perplexity) signals competition with established hyperscalers.
Custom AI chips from Nvidia, AMD, and others cater to diverse AI workloads, emphasizing a future where AI hardware becomes commoditized.
Diversification of AI Applications:
AI expanding into new domains like robotics (Google-Apptronik partnership), autonomous driving (Waymo in Tokyo), and small-scale generative AI experimentation (Nvidia Jetson Orin Nano).
Embodied AI and integration into real-world systems signal a broader trend toward physical and contextual AI solutions.
Accelerated AI Product Timelines:
Companies are speeding up product launches, particularly in reasoning models, despite concerns about readiness or diminishing returns (e.g., GPT-5 delays, Orion challenges).
AI hyperscalers are investing heavily in R&D while balancing resource constraints and competitive pressures.
Strategic AI Search Transformations:
Search is evolving toward conversational AI interfaces (e.g., Google’s AI Mode), moving away from traditional UX paradigms.
Real-time and personalized search capabilities (e.g., Perplexity AI) are disrupting established search norms.
AI Safety and Alignment:
A growing focus on AI safety, transparency, and ethical alignment, particularly in reasoning models and enterprise solutions.
Addressing concerns around hallucinations, misinformation, and operational risks through deliberate design and regulatory compliance.
Collaboration Between AI and Robotics:
Partnerships between AI and robotics companies (e.g., Apptronik and Google DeepMind) highlight the importance of combining software intelligence with physical systems.
Embodied AI is gaining traction for industrial, healthcare, and home applications.
With massive ♥️ Gennaro Cuofano, The Business Engineer
