AI is developing all the time. Here are some picks from several articles what is expected to happen in AI and around it in 2025. Here are picks from various articles, the texts are picks from the article edited and in some cases translated for clarity.
AI in 2025: Five Defining Themes
https://news.sap.com/2025/01/ai-in-2025-defining-themes/
Artificial intelligence (AI) is accelerating at an astonishing pace, quickly moving from emerging technologies to impacting how businesses run. From building AI agents to interacting with technology in ways that feel more like a natural conversation, AI technologies are poised to transform how we work.
But what exactly lies ahead?
1. Agentic AI: Goodbye Agent Washing, Welcome Multi-Agent Systems
AI agents are currently in their infancy. While many software vendors are releasing and labeling the first “AI agents” based on simple conversational document search, advanced AI agents that will be able to plan, reason, use tools, collaborate with humans and other agents, and iteratively reflect on progress until they achieve their objective are on the horizon. The year 2025 will see them rapidly evolve and act more autonomously. More specifically, 2025 will see AI agents deployed more readily “under the hood,” driving complex agentic workflows.
In short, AI will handle mundane, high-volume tasks while the value of human judgement, creativity, and quality outcomes will increase.
2. Models: No Context, No Value
Large language models (LLMs) will continue to become a commodity for vanilla generative AI tasks, a trend that has already started. LLMs are drawing on an increasingly tapped pool of public data scraped from the internet. This will only worsen, and companies must learn to adapt their models to unique, content-rich data sources.
We will also see a greater variety of foundation models that fulfill different purposes. Take, for example, physics-informed neural networks (PINNs), which generate outcomes based on predictions grounded in physical reality or robotics. PINNs are set to gain more importance in the job market because they will enable autonomous robots to navigate and execute tasks in the real world.
Models will increasingly become more multimodal, meaning an AI system can process information from various input types.
3. Adoption: From Buzz to Business
While 2024 was all about introducing AI use cases and their value for organizations and individuals alike, 2025 will see the industry’s unprecedented adoption of AI specifically for businesses. More people will understand when and how to use AI, and the technology will mature to the point where it can deal with critical business issues such as managing multi-national complexities. Many companies will also gain practical experience working for the first time through issues like AI-specific legal and data privacy terms (compared to when companies started moving to the cloud 10 years ago), building the foundation for applying the technology to business processes.
4. User Experience: AI Is Becoming the New UI
AI’s next frontier is seamlessly unifying people, data, and processes to amplify business outcomes. In 2025, we will see increased adoption of AI across the workforce as people discover the benefits of humans plus AI.
This means disrupting the classical user experience from system-led interactions to intent-based, people-led conversations with AI acting in the background. AI copilots will become the new UI for engaging with a system, making software more accessible and easier for people. AI won’t be limited to one app; it might even replace them one day. With AI, frontend, backend, browser, and apps are blurring. This is like giving your AI “arms, legs, and eyes.”
5. Regulation: Innovate, Then Regulate
It’s fair to say that governments worldwide are struggling to keep pace with the rapid advancements in AI technology and to develop meaningful regulatory frameworks that set appropriate guardrails for AI without compromising innovation.
12 AI predictions for 2025
This year we’ve seen AI move from pilots into production use cases. In 2025, they’ll expand into fully-scaled, enterprise-wide deployments.
https://www.cio.com/article/3630070/12-ai-predictions-for-2025.html
This year we’ve seen AI move from pilots into production use cases. In 2025, they’ll expand into fully-scaled, enterprise-wide deployments.
1. Small language models and edge computing
Most of the attention this year and last has been on the big language models — specifically on ChatGPT in its various permutations, as well as competitors like Anthropic’s Claude and Meta’s Llama models. But for many business use cases, LLMs are overkill and are too expensive, and too slow, for practical use.
“Looking ahead to 2025, I expect small language models, specifically custom models, to become a more common solution for many businesses,”
2. AI will approach human reasoning ability
In mid-September, OpenAI released a new series of models that thinks through problems much like a person would, it claims. The company says it can achieve PhD-level performance in challenging benchmark tests in physics, chemistry, and biology. For example, the previous best model, GPT-4o, could only solve 13% of the problems on the International Mathematics Olympiad, while the new reasoning model solved 83%.
If AI can reason better, then it will make it possible for AI agents to understand our intent, translate that into a series of steps, and do things on our behalf, says Gartner analyst Arun Chandrasekaran. “Reasoning also helps us use AI as more of a decision support system,”
3. Massive growth in proven use cases
This year, we’ve seen some use cases proven to have ROI, says Monteiro. In 2025, those use cases will see massive adoption, especially if the AI technology is integrated into the software platforms that companies are already using, making it very simple to adopt.
“The fields of customer service, marketing, and customer development are going to see massive adoption,”
4. The evolution of agile development
The agile manifesto was released in 2001 and, since then, the development philosophy has steadily gained over the previous waterfall style of software development.
“For the last 15 years or so, it’s been the de-facto standard for how modern software development works,”
5. Increased regulation
At the end of September, California governor Gavin Newsom signed a law requiring gen AI developers to disclose the data they used to train their systems, which applies to developers who make gen AI systems publicly available to Californians. Developers must comply by the start of 2026.
There are also regulations about the use of deep fakes, facial recognition, and more. The most comprehensive law, the EU’s AI Act, which went into effect last summer, is also something that companies will have to comply with starting in mid-2026, so, again, 2025 is the year when they will need to get ready.
6. AI will become accessible and ubiquitous
With gen AI, people are still at the stage of trying to figure out what gen AI is, how it works, and how to use it.
“There’s going to be a lot less of that,” he says. But gen AI will become ubiquitous and seamlessly woven into workflows, the way the internet is today.
7. Agents will begin replacing services
Software has evolved from big, monolithic systems running on mainframes, to desktop apps, to distributed, service-based architectures, web applications, and mobile apps. Now, it will evolve again, says Malhotra. “Agents are the next phase,” he says. Agents can be more loosely coupled than services, making these architectures more flexible, resilient and smart. And that will bring with it a completely new stack of tools and development processes.
8. The rise of agentic assistants
In addition to agents replacing software components, we’ll also see the rise of agentic assistants, adds Malhotra. Take for example that task of keeping up with regulations.
Today, consultants get continuing education to stay abreast of new laws, or reach out to colleagues who are already experts in them. It takes time for the new knowledge to disseminate and be fully absorbed by employees.
“But an AI agent can be instantly updated to ensure that all our work is compliant with the new laws,” says Malhotra. “This isn’t science fiction.”
9. Multi-agent systems
Sure, AI agents are interesting. But things are going to get really interesting when agents start talking to each other, says Babak Hodjat, CTO of AI at Cognizant. It won’t happen overnight, of course, and companies will need to be careful that these agentic systems don’t go off the rails.
Companies such as Sailes and Salesforce are already developing multi-agent workflows.
10. Multi-modal AI
Humans and the companies we build are multi-modal. We read and write text, we speak and listen, we see and we draw. And we do all these things through time, so we understand that some things come before other things. Today’s AI models are, for the most part, fragmentary. One can create images, another can only handle text, and some recent ones can understand or produce video.
11. Multi-model routing
Not to be confused with multi-modal AI, multi-modal routing is when companies use more than one LLM to power their gen AI applications. Different AI models are better at different things, and some are cheaper than others, or have lower latency. And then there’s the matter of having all your eggs in one basket.
“A number of CIOs I’ve spoken with recently are thinking about the old ERP days of vendor lock,” says Brett Barton, global AI practice leader at Unisys. “And it’s top of mind for many as they look at their application portfolio, specifically as it relates to cloud and AI capabilities.”
Diversifying away from using just a single model for all use cases means a company is less dependent on any one provider and can be more flexible as circumstances change.
12. Mass customization of enterprise software
Today, only the largest companies, with the deepest pockets, get to have custom software developed specifically for them. It’s just not economically feasible to build large systems for small use cases.
“Right now, people are all using the same version of Teams or Slack or what have you,” says Ernst & Young’s Malhotra. “Microsoft can’t make a custom version just for me.” But once AI begins to accelerate the speed of software development while reducing costs, it starts to become much more feasible.
9 IT resolutions for 2025
https://www.cio.com/article/3629833/9-it-resolutions-for-2025.html
1. Innovate
“We’re embracing innovation,”
2. Double down on harnessing the power of AI
Not surprisingly, getting more out of AI is top of mind for many CIOs.
“I am excited about the potential of generative AI, particularly in the security space,”
3. And ensure effective and secure AI rollouts
“AI is everywhere, and while its benefits are extensive, implementing it effectively across a corporation presents challenges. Balancing the rollout with proper training, adoption, and careful measurement of costs and benefits is essential, particularly while securing company assets in tandem,”
4. Focus on responsible AI
The possibilities of AI grow by the day — but so do the risks.
“My resolution is to mature in our execution of responsible AI,”
“AI is the new gold and in order to truly maximize it’s potential, we must first have the proper guardrails in place. Taking a human-first approach to AI will help ensure our state can maintain ethics while taking advantage of the new AI innovations.”
5. Deliver value from generative AI
As organizations move from experimenting and testing generative AI use cases, they’re looking for gen AI to deliver real business value.
“As we go into 2025, we’ll continue to see the evolution of gen AI. But it’s no longer about just standing it up. It’s more about optimizing and maximizing the value we’re getting out of gen AI,”
6. Empower global talent
Although harnessing AI is a top objective for Morgan Stanley’s Wetmur, she says she’s equally committed to harnessing the power of people.
7. Create a wholistic learning culture
Wetmur has another talent-related objective: to create a learning culture — not just in her own department but across all divisions.
8. Deliver better digital experiences
Deltek’s Cilsick has her sights set on improving her company’s digital employee experience, believing that a better DEX will yield benefits in multiple ways.
Cilsick says she first wants to bring in new technologies and automation to “make things as easy as possible,” mirroring the digital experiences most workers have when using consumer technologies.
“It’s really about leveraging tech to make sure [employees] are more efficient and productive,”
“In 2025 my primary focus as CIO will be on transforming operational efficiency, maximizing business productivity, and enhancing employee experiences,”
9. Position the company for long-term success
Lieberman wants to look beyond 2025, saying another resolution for the year is “to develop a longer-term view of our technology roadmap so that we can strategically decide where to invest our resources.”
“My resolutions for 2025 reflect the evolving needs of our organization, the opportunities presented by AI and emerging technologies, and the necessity to balance innovation with operational efficiency,”
Lieberman aims to develop AI capabilities to automate routine tasks.
“Bots will handle common inquiries ranging from sales account summaries to HR benefits, reducing response times and freeing up resources for strategic initiatives,”
Not just hype — here are real-world use cases for AI agents
https://venturebeat.com/ai/not-just-hype-here-are-real-world-use-cases-for-ai-agents/
Just seven or eight months ago, when a customer called in to or emailed Baca Systems with a service question, a human agent handling the query would begin searching for similar cases in the system and analyzing technical documents.
This process would take roughly five to seven minutes; then the agent could offer the “first meaningful response” and finally begin troubleshooting.
But now, with AI agents powered by Salesforce, that time has been shortened to as few as five to 10 seconds.
Now, instead of having to sift through databases for previous customer calls and similar cases, human reps can ask the AI agent to find the relevant information. The AI runs in the background and allows humans to respond right away, Russo noted.
AI can serve as a sales development representative (SDR) to send out general inquires and emails, have a back-and-forth dialogue, then pass the prospect to a member of the sales team, Russo explained.
But once the company implements Salesforce’s Agentforce, a customer needing to modify an order will be able to communicate their needs with AI in natural language, and the AI agent will automatically make adjustments. When more complex issues come up — such as a reconfiguration of an order or an all-out venue change — the AI agent will quickly push the matter up to a human rep.
Open Source in 2025: Strap In, Disruption Straight Ahead
Look for new tensions to arise in the New Year over licensing, the open source AI definition, security and compliance, and how to pay volunteer maintainers.
https://thenewstack.io/open-source-in-2025-strap-in-disruption-straight-ahead/
The trend of widely used open source software moving to more restrictive licensing isn’t new.
In addition to the demands of late-stage capitalism and impatient investors in companies built on open source tools, other outside factors are pressuring the open source world. There’s the promise/threat of generative AI, for instance. Or the shifting geopolitical landscape, which brings new security concerns and governance regulations.
What’s ahead for open source in 2025?
More Consolidation, More Licensing Changes
The Open Source AI Debate: Just Getting Started
Security and Compliance Concerns Will Rise
Paying Maintainers: More Cash, Creativity Needed
Kyberturvallisuuden ja tekoälyn tärkeimmät trendit 2025
https://www.uusiteknologia.fi/2024/11/20/kyberturvallisuuden-ja-tekoalyn-tarkeimmat-trendit-2025/
1. Cyber infrastructure will be centered on a single, unified security platform
2. Big data will give an edge against new entrants
3. AI’s integrated role in 2025 means building trust, governance engagement, and a new kind of leadership
4. Businesses will adopt secure enterprise browsers more widely
5. AI’s energy implications will be more widely recognized in 2025
6. Quantum realities will become clearer in 2025
7. Security and marketing leaders will work more closely together
Presentation: For 2025, ‘AI eats the world’.
https://www.ben-evans.com/presentations
Just like other technologies that have gone before, such as cloud and cybersecurity automation, right now AI lacks maturity.
https://www.securityweek.com/ai-implementing-the-right-technology-for-the-right-use-case/
If 2023 and 2024 were the years of exploration, hype and excitement around AI, 2025 (and 2026) will be the year(s) that organizations start to focus on specific use cases for the most productive implementations of AI and, more importantly, to understand how to implement guardrails and governance so that it is viewed as less of a risk by security teams and more of a benefit to the organization.
Businesses are developing applications that add Large Language Model (LLM) capabilities to provide superior functionality and advanced personalization
Employees are using third party GenAI tools for research and productivity purposes
Developers are leveraging AI-powered code assistants to code faster and meet challenging production deadlines
Companies are building their own LLMs for internal use cases and commercial purposes.
AI is still maturing
However, just like other technologies that have gone before, such as cloud and cybersecurity automation, right now AI lacks maturity. Right now, we very much see AI in this “peak of inflated expectations” phase and predict that it will dip into the “trough of disillusionment”, where organizations realize that it is not the silver bullet they thought it would be. In fact, there are already signs of cynicism as decision-makers are bombarded with marketing messages from vendors and struggle to discern what is a genuine use case and what is not relevant for their organization.
There is also regulation that will come into force, such as the EU AI Act, which is a comprehensive legal framework that sets out rules for the development and use of AI.
AI certainly won’t solve every problem, and it should be used like automation, as part of a collaborative mix of people, process and technology. You simply can’t replace human intuition with AI, and many new AI regulations stipulate that human oversight is maintained.
7 Splunk Predictions for 2025
https://www.splunk.com/en_us/form/future-predictions.html
AI: Projects must prove their worth to anxious boards or risk defunding, and LLMs will go small to reduce operating costs and environmental impact.
OpenAI, Google and Anthropic Are Struggling to Build More Advanced AI
Three of the leading artificial intelligence companies are seeing diminishing returns from their costly efforts to develop newer models.
https://www.bloomberg.com/news/articles/2024-11-13/openai-google-and-anthropic-are-struggling-to-build-more-advanced-ai
Sources: OpenAI, Google, and Anthropic are all seeing diminishing returns from costly efforts to build new AI models; a new Gemini model misses internal targets
It Costs So Much to Run ChatGPT That OpenAI Is Losing Money on $200 ChatGPT Pro Subscriptions
https://futurism.com/the-byte/openai-chatgpt-pro-subscription-losing-money?fbclid=IwY2xjawH8epVleHRuA2FlbQIxMQABHeggEpKe8ZQfjtPRC0f2pOI7A3z9LFtFon8lVG2VAbj178dkxSQbX_2CJQ_aem_N_ll3ETcuQ4OTRrShHqNGg
In a post on X-formerly-Twitter, CEO Sam Altman admitted an “insane” fact: that the company is “currently losing money” on ChatGPT Pro subscriptions, which run $200 per month and give users access to its suite of products including its o1 “reasoning” model.
“People use it much more than we expected,” the cofounder wrote, later adding in response to another user that he “personally chose the price and thought we would make some money.”
Though Altman didn’t explicitly say why OpenAI is losing money on these premium subscriptions, the issue almost certainly comes down to the enormous expense of running AI infrastructure: the massive and increasing amounts of electricity needed to power the facilities that power AI, not to mention the cost of building and maintaining those data centers. Nowadays, a single query on the company’s most advanced models can cost a staggering $1,000.
Tekoäly edellyttää yhä nopeampia verkkoja
https://etn.fi/index.php/opinion/16974-tekoaely-edellyttaeae-yhae-nopeampia-verkkoja
A resilient digital infrastructure is critical to effectively harnessing telecommunications networks for AI innovations and cloud-based services. The increasing demand for data-rich applications related to AI requires a telecommunications network that can handle large amounts of data with low latency, writes Carl Hansson, Partner Solutions Manager at Orange Business.
AI’s Slowdown Is Everyone Else’s Opportunity
Businesses will benefit from some much-needed breathing space to figure out how to deliver that all-important return on investment.
https://www.bloomberg.com/opinion/articles/2024-11-20/ai-slowdown-is-everyone-else-s-opportunity
Näin sirumarkkinoilla käy ensi vuonna
https://etn.fi/index.php/13-news/16984-naein-sirumarkkinoilla-kaey-ensi-vuonna
The growing demand for high-performance computing (HPC) for artificial intelligence and HPC computing continues to be strong, with the market set to grow by more than 15 percent in 2025, IDC estimates in its recent Worldwide Semiconductor Technology Supply Chain Intelligence report.
IDC predicts eight significant trends for the chip market by 2025.
1. AI growth accelerates
2. Asia-Pacific IC Design Heats Up
3. TSMC’s leadership position is strengthening
4. The expansion of advanced processes is accelerating.
5. Mature process market recovers
6. 2nm Technology Breakthrough
7. Restructuring the Packaging and Testing Market
8. Advanced packaging technologies on the rise
2024: The year when MCUs became AI-enabled
https://www-edn-com.translate.goog/2024-the-year-when-mcus-became-ai-enabled/?fbclid=IwZXh0bgNhZW0CMTEAAR1_fEakArfPtgGZfjd-NiPd_MLBiuHyp9qfiszczOENPGPg38wzl9KOLrQ_aem_rLmf2vF2kjDIFGWzRVZWKw&_x_tr_sl=en&_x_tr_tl=fi&_x_tr_hl=fi&_x_tr_pto=wapp
The AI party in the MCU space started in 2024, and in 2025, it is very likely that there will be more advancements in MCUs using lightweight AI models.
Adoption of AI acceleration features is a big step in the development of microcontrollers. The inclusion of AI features in microcontrollers started in 2024, and it is very likely that in 2025, their features and tools will develop further.
Just like other technologies that have gone before, such as cloud and cybersecurity automation, right now AI lacks maturity.
https://www.securityweek.com/ai-implementing-the-right-technology-for-the-right-use-case/
If 2023 and 2024 were the years of exploration, hype and excitement around AI, 2025 (and 2026) will be the year(s) that organizations start to focus on specific use cases for the most productive implementations of AI and, more importantly, to understand how to implement guardrails and governance so that it is viewed as less of a risk by security teams and more of a benefit to the organization.
Businesses are developing applications that add Large Language Model (LLM) capabilities to provide superior functionality and advanced personalization
Employees are using third party GenAI tools for research and productivity purposes
Developers are leveraging AI-powered code assistants to code faster and meet challenging production deadlines
Companies are building their own LLMs for internal use cases and commercial purposes.
AI is still maturing
AI Regulation Gets Serious in 2025 – Is Your Organization Ready?
While the challenges are significant, organizations have an opportunity to build scalable AI governance frameworks that ensure compliance while enabling responsible AI innovation.
https://www.securityweek.com/ai-regulation-gets-serious-in-2025-is-your-organization-ready/
Similar to the GDPR, the EU AI Act will take a phased approach to implementation. The first milestone arrives on February 2, 2025, when organizations operating in the EU must ensure that employees involved in AI use, deployment, or oversight possess adequate AI literacy. Thereafter from August 1 any new AI models based on GPAI standards must be fully compliant with the act. Also similar to GDPR is the threat of huge fines for non-compliance – EUR 35 million or 7 percent of worldwide annual turnover, whichever is higher.
While this requirement may appear manageable on the surface, many organizations are still in the early stages of defining and formalizing their AI usage policies.
Later phases of the EU AI Act, expected in late 2025 and into 2026, will introduce stricter requirements around prohibited and high-risk AI applications. For organizations, this will surface a significant governance challenge: maintaining visibility and control over AI assets.
Tracking the usage of standalone generative AI tools, such as ChatGPT or Claude, is relatively straightforward. However, the challenge intensifies when dealing with SaaS platforms that integrate AI functionalities on the backend. Analysts, including Gartner, refer to this as “embedded AI,” and its proliferation makes maintaining accurate AI asset inventories increasingly complex.
Where frameworks like the EU AI Act grow more complex is their focus on ‘high-risk’ use cases. Compliance will require organizations to move beyond merely identifying AI tools in use; they must also assess how these tools are used, what data is being shared, and what tasks the AI is performing. For instance, an employee using a generative AI tool to summarize sensitive internal documents introduces very different risks than someone using the same tool to draft marketing content.
For security and compliance leaders, the EU AI Act represents just one piece of a broader AI governance puzzle that will dominate 2025.
The next 12-18 months will require sustained focus and collaboration across security, compliance, and technology teams to stay ahead of these developments.
The Global Partnership on Artificial Intelligence (GPAI) is a multi-stakeholder initiative which aims to bridge the gap between theory and practice on AI by supporting cutting-edge research and applied activities on AI-related priorities.
https://gpai.ai/about/#:~:text=The%20Global%20Partnership%20on%20Artificial,activities%20on%20AI%2Drelated%20priorities.
830 Comments
Tomi Engdahl says:
https://huggingface.co/NousResearch/DeepHermes-3-Llama-3-8B-Preview-GGUF
Tomi Engdahl says:
Brain-inspired Computing Is Ready for the Big Time Neuromorphic pioneer Steve Furber says it’s just awaiting a killer app
https://spectrum.ieee.org/neuromorphic-computing-2671121824
Tomi Engdahl says:
AI Axe Falls On Tech, Anthropic’s Lightspeed Raise, Firefly Video, Anduril Takes On Army’s XR Contract
https://www.forbes.com/sites/charliefink/2025/02/13/ai-axe-falls-on-tech-anthropics-lightspeed-raise-firefly-video-anduril-takes-on-armys-xr-contract/
AI Axing so-called ‘underperformers’ at Google and Meta. The pain is spread worldwide and will affect several thousand employees, many of whom earn mid-six figure salaries. While technology jobs are among the first impacted by AI-powered workforce reduction, these workers still have their pick of repalcement jobs. AI makes good programmers great programmers. And great programmers do the work of two or more people.
Tomi Engdahl says:
Buoyed by AI, Cisco sees ‘lots’ of telcos planning edge rollouts
Cisco has raised full-year guidance as it prepares to take Ethernet technology into AI data centers.
https://www.lightreading.com/ai-machine-learning/buoyed-by-ai-cisco-sees-lots-of-telcos-planning-edge-rollouts
Tomi Engdahl says:
GitHub Copilot now available for free in mobile and the CLI
https://github.blog/changelog/2025-02-12-github-copilot-chat-and-github-copilot-extension-now-available-for-free-on-github-mobile-and-github-cli/
GitHub Copilot Chat in GitHub Mobile and Copilot Extension for the GitHub CLI are now available for free!
GitHub Copilot Chat on GitHub Mobile
Whether you’re tackling coding questions, brainstorming ideas, or working on the go, GitHub Copilot Chat is here to make collaboration faster and easier, no matter where you are.
On mobile, simply sign in with your personal GitHub account and tap the Copilot button to access 2,000 code completions and 50 chat messages per month! If you reach your quota, you can upgrade through an in-app purchase to enjoy unlimited access.
Download or update GitHub Mobile apps today from the Apple App Store or Google Play Store to experience the AI coding assistance right at your fingertips.
Tomi Engdahl says:
Anthropic prepares new Claude hybrid LLMs with reasoning capability
https://the-decoder.com/anthropic-prepares-new-claude-hybrid-llms-with-reasoning-capability/
Anthropic is getting ready to release a new AI model that combines traditional language model capabilities with advanced reasoning functions.
Tomi Engdahl says:
How Developers Are Using Bolt, a Fast Growing AI Coding Tool
StackBlitz was a browser-based IDE, but it’s now pivoted to an agentic IDE called Bolt. We talk to its CEO about how developers are using it.
https://thenewstack.io/how-developers-are-using-bolt-a-fast-growing-ai-coding-tool/
Tomi Engdahl says:
Dual-domain architecture shows almost 40 times higher energy efficiency for running neural networks
https://techxplore.com/news/2025-02-dual-domain-architecture-higher-energy.html
Tomi Engdahl says:
https://www.visualcapitalist.com/cp/visualizing-the-current-generation-of-humanoid-robots/
Tomi Engdahl says:
https://dev.to/devteam/the-future-of-4ml4
Hey folks! We wanted to let you know about our new subforem called Future. This is a dedicated space for talking about cutting-edge technologies that influence our careers, our day-to-day lives, culture, and everything of that nature.
https://future.forem.com/
Tomi Engdahl says:
People are squashing DeepSeek onto their Raspberry Pi mere days after it hit the public eye
https://www.xda-developers.com/deepseek-raspberry-pi-mere-days/
Summary
DeepSeek is an LLM from China that’s causing turbulence in the tech market.
An older version of DeepSeek can run on Raspberry Pi 5 but with slow performance.
The Raspberry Pi community may find a way to enhance DeepSeek’s performance.
Tomi Engdahl says:
https://www.uusiteknologia.fi/2024/12/18/konenako-helpottaa-kuljetusrobottien-toimintaa/
Tomi Engdahl says:
However, it’s not just DeepSeek’s latest AI. Meta’s open-source Llama 3.1 model also flunked almost as badly as DeepSeek’s R1 in a comparison test, with a 96 percent attack success rate (compared to dismal 100 percent for DeepSeek).
OpenAI’s recently released reasoning model, o1-preview, fared much better, with an attack success rate of just 26 percent.
In short, DeepSeek’s flaws deserve plenty of scrutiny going forward.
“DeepSeek is just another example of how every model can be broken — it’s just a matter of how much effort you put in,” Adversa AI CEO Alex Polyakov told Wired. “If you’re not continuously red-teaming your AI, you’re already compromised.”
https://futurism.com/deepseek-failed-every-security-test?fbclid=IwY2xjawIfnsVleHRuA2FlbQIxMQABHeDMorxfW0a1evXWweQe7qSwTrFJ8ODF5cZXqTqkL9zTsqmBpAdDWrHNtQ_aem__2_W2IE1Tq26yaYGQGK8MQ
Tomi Engdahl says:
https://futurism.com/allstate-demanding-delete-quotes-ai
Tomi Engdahl says:
https://futurism.com/openai-o3-release
Tomi Engdahl says:
Suno AI killed music industry
https://youtu.be/UAHaCyw7rNA?si=17dNVCaGrzvnn78K
Tomi Engdahl says:
Mark Gurman / Bloomberg:
A look at Apple’s and Meta’s robotics efforts; sources say Apple is working on a 27″ Studio Display successor to launch by 2026 and considering monetizing Maps — Apple and Meta are set to go toe-to-toe in another emerging area: AI-powered humanoid robots.
https://www.bloomberg.com/news/newsletters/2025-02-16/apple-and-meta-are-set-to-battle-over-new-area-humanoid-robots-m77mwid3?accessToken=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJzb3VyY2UiOiJTdWJzY3JpYmVyR2lmdGVkQXJ0aWNsZSIsImlhdCI6MTczOTc1MDEyOCwiZXhwIjoxNzQwMzU0OTI4LCJhcnRpY2xlSWQiOiJTUlMxR0dEV0xVNjgwMCIsImJjb25uZWN0SWQiOiIzOEVGMDZFMzkwMzc0NkEyQkZFODVERURGRDMwQ0UzOCJ9.q4iH9-HhPWTlu7daXRbMhF2wUL4B8278GcqyE1ayeD0
Tomi Engdahl says:
Max Tani / Semafor:
Documents: the NYT greenlights using select internal and external AI tools for editorial and product staff, including from OpenAI, Amazon, Google, and Microsoft — The Scoop — The New York Times is greenlighting the use of AI for its product and editorial staff, saying that internal tools …
New York Times goes all-in on internal AI tools
https://www.semafor.com/article/02/16/2025/new-york-times-goes-all-in-on-internal-ai-tools
The New York Times is greenlighting the use of AI for its product and editorial staff, saying that internal tools could eventually write social copy, SEO headlines, and some code.
In an email to newsroom staff, the company announced that it’s opening up AI training to the newsroom, and debuting a new internal AI tool called Echo to staff, Semafor has learned. The Times also shared documents and videos laying out editorial do’s and don’t for using AI, and shared a suite of AI products that staff could now use to develop web products and editorial ideas.
“Generative AI can assist our journalists in uncovering the truth and helping more people understand the world. Machine learning already helps us report stories we couldn’t otherwise, and generative AI has the potential to bolster our journalistic capabilities even more,” the company’s editorial guidelines said.
Tomi Engdahl says:
Namanyay Goel / N’s Blog:
New junior developers are reliant on Copilot, Claude, and other AI tools, meaning foundational coding knowledge is now missing, creating problems down the road
New Junior Developers Can’t Actually Code
https://nmn.gl/blog/ai-and-learning
Feb 14 2025
Something’s been bugging me about how new devs and I need to talk about it.
We’re at this weird inflection point in software development. Every junior dev I talk to has Copilot or Claude or GPT running 24/7. They’re shipping code faster than ever. But when I dig deeper into their understanding of what they’re shipping? That’s where things get concerning.
Sure, the code works, but ask why it works that way instead of another way? Crickets. Ask about edge cases? Blank stares.
The foundational knowledge that used to come from struggling through problems is just… missing.
We’re trading deep understanding for quick fixes, and while it feels great in the moment, we’re going to pay for this later.
Back when we had to actually think
I recently realized that there’s a whole generation of new programmers who don’t even know what StackOverflow is.
Back when “Claude” was not a chatbot but the man who invented the field of information entropy, there was a different way to debug programming problems.
First, search on Google. Then, hope some desperate soul had posed a similar question as you had. If they did, you’d find a detailed, thoughtful, (and often patronizing) answer from a wise greybeard on this site called “Stack Overflow”.
Tomi Engdahl says:
Dan Gallagher / Wall Street Journal:
Intel’s foundry business lost $13B+ on $17.5B in revenue in 2024, while TSMC generated $41.1B in operating profit on $90B in revenue over the same period
Intel’s Stock Has Soared, but a Rescue Will Be Hard to Pull Off
While President Trump’s domestic AI focus is a good sign for Intel, easy fixes have proved elusive
https://www.wsj.com/tech/intels-stock-is-soaring-but-a-rescue-will-be-hard-to-pull-off-0a19552c?st=DvHD4q&reflink=desktopwebshare_permalink
Broadcom, TSMC Weigh Possible Intel Deals That Would Split Storied Chip Maker
Broadcom has interest in Intel’s chip-design business, while TSMC is looking at the company’s factories
https://www.wsj.com/tech/broadcom-tsmc-eye-possible-intel-deals-that-would-split-storied-chip-maker-966b143b?st=RiRV14&reflink=desktopwebshare_permalink
Tomi Engdahl says:
Reuters:
South Korea’s data protection authority says DeepSeek app downloads have been suspended in the country after DeepSeek failed to follow its personal data rules
New downloads of DeepSeek suspended in South Korea, data protection agency says
https://www.reuters.com/technology/south-koreas-data-protection-authority-suspends-local-service-deepseek-2025-02-17/
SEOUL, Feb 17 (Reuters) – South Korea’s data protection authority on Monday said new downloads of the Chinese AI app DeepSeek had been suspended in the country after DeepSeek acknowledged failing to take into account some of the agency’s rules on protecting personal data.
The service of the app will be resumed once improvements are made in accordance with the country’s privacy law, the Personal Information Protection Commission (PIPC) said in a media briefing.
Tomi Engdahl says:
Paul Sawers / TechCrunch:
A look at the OpenEuroLLM project, a partnership among 20 EU organizations to develop open-source LLMs that support all EU languages, with a budget of €37.4M — Large language models (LLMs) landed on Europe’s digital sovereignty agenda with a bang last week, as news emerged of a new program …
Open source LLMs hit Europe’s digital sovereignty roadmap
https://techcrunch.com/2025/02/16/open-source-llms-hit-europes-digital-sovereignty-roadmap/
Large language models (LLMs) landed on Europe’s digital sovereignty agenda with a bang last week, as news emerged of a new program to develop a series of “truly” open source LLMs covering all European Union languages.
This includes the current 24 official EU languages, as well as languages for countries currently negotiating for entry to the EU market, such as Albania. Future-proofing is the name of the game.
OpenEuroLLM is a collaboration between some 20 organizations, co-led by Jan Hajič, a computational linguist from the Charles University in Prague, and Peter Sarlin, CEO and co-founder of Finnish AI lab Silo AI, which AMD acquired last year for $665 million.
The project fits a broader narrative that has seen Europe push digital sovereignty as a priority, enabling it to bring mission-critical infrastructure and tools closer to home. Most of the cloud giants are investing in local infrastructure to ensure EU data stays local, while AI darling OpenAI recently unveiled a new offering that allows customers to process and store data in Europe.
Elsewhere, the EU recently signed an $11 billion deal to create a sovereign satellite constellation to rival Elon Musk’s Starlink.
So OpenEuroLLM is certainly on-brand.
Tomi Engdahl says:
Financial Times:
Match Head of Trust and Safety Yoel Roth says Match is using AI in its apps, like Tinder, to detect men’s “off-color” messages and drive “behavioral change”
https://www.ft.com/content/4e39d08b-41ef-41ea-abc0-952d06324484
Tomi Engdahl says:
Reuters:
Tencent is testing using DeepSeek for search in its messaging app Weixin, while Baidu plans to fully connect its search engine to DeepSeek and its Ernie LLM
Tencent’s Weixin app, Baidu launch DeepSeek search testing
https://www.reuters.com/technology/artificial-intelligence/tencents-messaging-app-weixin-launches-beta-testing-with-deepseek-2025-02-16/
HONG KONG/SHANGHAI, Feb 16 (Reuters) – Tencent said on Sunday its (0700.HK)
, opens new tab Weixin messaging app, China’s largest, is allowing some users to search via DeepSeek’s artificial intelligence model as firms race to link up with the AI startup.
In a beta test, Weixin is testing access to DeepSeek for searches, Tencent said in a statement to Reuters.
The move by the Chinese tech giant is notable as integrating DeepSeek brings in an external AI platform, while tech firms compete fiercely in developing the most advanced AI.
Tomi Engdahl says:
Financial Times:
How Walmart’s tech investments, including an AI tool to plan worker shifts, helped it take on Amazon, with 18% of its ~$680B revenue in FY 2024 generated online
How a resurgent Walmart saw off the Amazon threat
A decade ago, its huge stores looked outdated as online sales grew. Now they are pulling in more customers than ever
https://www.ft.com/content/13e6ba39-8ef8-4ac1-9079-d8e7ec53d3c2?accessToken=zwAGLindhEFokc8T5ro5jvhKwdOQedjn7FPTwg.MEUCIEYbXQq_B7kglc1ZV7G4M63gDW4E8EXdCPMrqv-TDeklAiEAt6oFkgy4afLytSteJfEczx7AlIQyyS5YJgzDKyEhdJM&sharetype=gift&token=5f594eb7-b8fa-40f5-9df6-95cd6d122813
A neighbouring Chanel boutique in the city’s SoHo shopping district was selling denim jackets for $4,400. But the pop-up was offering $38 faux leather jackets and jeans for $26. The owner was not a designer label; it was Walmart.
“We have been on this mission to democratise fashion,” Denise Incandela, executive vice-president for fashion at Walmart US, told the stylish crowd last week. “I hope you feel that we’re making progress based on the prices that you see today and the incredible quality.”
Less than a decade ago, investors feared for the group’s future as ecommerce sales grew rapidly. In 2015 Amazon overtook Walmart’s market capitalisation, its slick delivery services making huge stores on the edges of towns seem anachronistic.
Many expected Amazon’s 2017 acquisition of Whole Foods to presage an assault on the US grocery market. In the year to January 2019, Walmart reported its lowest net income since fiscal 2002.
Today, the company founded by Sam Walton 63 years ago is resurgent.
Analysts expect it to report a record $681bn of revenue when it releases full-year results on February 20, maintaining its status as the world’s largest company by sales.
the investments it has made in online infrastructure are paying off.
Ecommerce sales have been expanding by more than 20 per cent a year. Group-wide, 18 per cent of revenue is now generated online and its marketplace lists more than 700mn items from third-party merchants.
“Walmart is rapidly transforming itself into a tech company akin to Amazon,”
Half of the recent growth in US retail sales has been absorbed by just three companies: Walmart, Amazon and the warehouse club chain Costco, according to Morgan Stanley.
Among the top priorities is automating its stores and warehouses. All 42 of Walmart’s US regional distribution centres are being fitted out with robots that assemble pallets categorised by department to speed up stocking when they arrive in stores.
In the company’s vast Supercenters, clerks scan stacks of inventory with their phones to learn which shelves must be restocked and where to find supplies in storerooms. A phone-based artificial intelligence tool enables managers to plan worker shifts in five minutes, a task that once took an hour, while digital shelf-edge labelling will end the chore of manually updating price tags.
“We want to be the low-cost provider. We want customers to think of us as the place to go to get the lowest prices on anything that they want to buy,” says John David Rainey, chief financial officer. “To do that, we have to have the lowest cost to serve.”
Tomi Engdahl says:
Downloads of DeepSeek’s AI Apps Paused in South Korea Over Privacy Concerns
https://www.securityweek.com/downloads-of-deepseeks-ai-apps-paused-in-south-korea-over-privacy-concerns/
DeepSeek has temporarily paused downloads of its chatbot apps in South Korea while it works with local authorities to address privacy concerns.
Tomi Engdahl says:
USB Stick Hides Large Language Model
https://hackaday.com/2025/02/17/usb-stick-hides-large-language-model/
Large language models (LLMs) are all the rage in the generative AI world these days, with the truly large ones like GPT, LLaMA, and others using tens or even hundreds of billions of parameters to churn out their text-based responses. These typically require glacier-melting amounts of computing hardware, but the “large” in “large language models” doesn’t really need to be that big for there to be a functional, useful model. LLMs designed for limited hardware or consumer-grade PCs are available now as well, but [Binh] wanted something even smaller and more portable, so he put an LLM on a USB stick.
This USB stick isn’t just a jump drive with a bit of memory on it, though. Inside the custom 3D printed case is a Raspberry Pi Zero W running llama.cpp, a lightweight, high-performance version of LLaMA. Getting it on this Pi wasn’t straightforward at all, though, as the latest version of llama.cpp is meant for ARMv8 and this particular Pi was running the ARMv6 instruction set. That meant that [Binh] needed to change the source code to remove the optimizations for the more modern ARM machines, but with a week’s worth of effort spent on it he finally got the model on the older Raspberry Pi.
World’s First USB Stick with Local LLM – AI in Your Pocket!
https://www.youtube.com/watch?v=SM-fFsE9EDU
Cherry on top of the cake, it requires no dependency, you can connect it to any computer, create a new file and the content will be automatically generated from the USB side. Essentially the first ever native LLM USB.
This was done on an 8-year-old pi zero, which has 512MB of Ram and an arm1176jzf-s CPU.
To be able to run LLM, let alone with llama.cpp on this was quite something. Arm1176jzf-s was first released in 2002, it implements armv6l isa. It took 12 hours just to compile the whole source of llamacpp and more than a week for me to make it run on an unsupported isa.
The performance is quite terrible and offer no practical use, but it is a fun look into the future, where LLM can run potentially anywhere.
00:00 – Intro
00:20 – Hardware & Casing
01:48 – Case Assembly
02:17 – Using Llama.cpp
02:51 – Fixing Llama.cpp
05:14 – LLM Demo & Benchmark
07:30 – Building a real USB
09:30 – USB Demo
11:57 – Endnote
Tomi Engdahl says:
Large Language Models On Small Computers
https://hackaday.com/2024/09/07/large-language-models-on-small-computers/
As technology progresses, we generally expect processing capabilities to scale up. Every year, we get more processor power, faster speeds, greater memory, and lower cost. However, we can also use improvements in software to get things running on what might otherwise be considered inadequate hardware. Taking this to the extreme, while large language models (LLMs) like GPT are running out of data to train on and having difficulty scaling up, [DaveBben] is experimenting with scaling down instead, running an LLM on the smallest computer that could reasonably run one.
Of course, some concessions have to be made to get an LLM running on underpowered hardware. In this case, the computer of choice is an ESP32, so the dataset was reduced from the trillions of parameters of something like GPT-4 or even hundreds of billions for GPT-3 down to only 260,000. The dataset comes from the tinyllamas checkpoint, and llama.2c is the implementation that [DaveBben] chose for this setup, as it can be streamlined to run a bit better on something like the ESP32. The specific model is the ESP32-S3FH4R2, which was chosen for its large amount of RAM compared to other versions since even this small model needs a minimum of 1 MB to run. It also has two cores, which will both work as hard as possible under (relatively) heavy loads like these, and the clock speed of the CPU can be maxed out at around 240 MHz.
https://github.com/DaveBben/esp32-llm
Tomi Engdahl says:
World’s First USB Stick with Local LLM – AI in Your Pocket!
https://www.youtube.com/watch?v=SM-fFsE9EDU
I created the first USB in the world that has LLM running locally on it.
Cherry on top of the cake, it requires no dependency, you can connect it to any computer, create a new file and the content will be automatically generated from the USB side. Essentially the first ever native LLM USB.
This was done on an 8-year-old pi zero, which has 512MB of Ram and an arm1176jzf-s CPU.
To be able to run LLM, let alone with llama.cpp on this was quite something. Arm1176jzf-s was first released in 2002, it implements armv6l isa. It took 12 hours just to compile the whole source of llamacpp and more than a week for me to make it run on an unsupported isa.
The performance is quite terrible and offer no practical use, but it is a fun look into the future, where LLM can run potentially anywhere.
Benchmark of models
Tiny-15M: 223ms/token
Lamini-T5-Flan-77M: 2.5s/token
SmolLM2-136M: 2.2s/token
Chapters
00:00 – Intro
00:20 – Hardware & Casing
01:48 – Case Assembly
02:17 – Using Llama.cpp
02:51 – Fixing Llama.cpp
05:14 – LLM Demo & Benchmark
07:30 – Building a real USB
09:30 – USB Demo
11:57 – Endnote
Tomi Engdahl says:
Digiuutiset
Näiden laitteiden piti mullistaa markkinat – Mitä sitten tapahtui?
Tekoälyä hyödyntävien tietokoneiden oli ennustettu elvyttävän PC-markkinoita, mutta niiden läpimurto on viivästynyt.
https://www.iltalehti.fi/digiuutiset/a/e38b4914-e4d3-4a26-98df-f5ed98bb32d1
Tekoälyä hyödyntävistä tietokoneista on odotettu piristystä PC-markkinoille, mutta toistaiseksi niiden läpimurto on viivästynyt. Markkinatutkimusyhtiö Gartner ennusti viime vuonna, että tekoälymallit valloittavat markkinat vuoteen 2026 mennessä.
Tuoreiden Contextin markkinatietojen perusteella kehitys on kuitenkin hitaampaa kuin odotettiin. Yrityksen keräämien myyntilukujen mukaan loka–joulukuussa 2024 Euroopassa myydyistä tietokoneista 40 prosenttia oli tekoälymalleja, mutta se ei vielä tarkoita merkittävää murrosta alalla, kirjoittaa teknologiasivusto The Register.
– Valmistajat lisäävät tekoälytoiminnallisuuksia yhä useampiin laitteisiin, mikä tekee tekoäly-PC:stä väistämättömän kehityksen suunnan, eikä niinkään kuluttajien valinnan. Tämä ei kuitenkaan tarkoita, että kuluttajat aktiivisesti etsisivät näitä ominaisuuksia, Contextin vanhempi analyytikko Marie-Christine Pygott sanoo The Registerin mukaan.
Tekoälyä hyödyntävistä tietokoneista vain viisi prosenttia kuului Microsoftin Copilot+ -mallistoon, joka edustaa Windows 11 -käyttöjärjestelmällä toimivia, tekoälyoptimoituja PC-laitteita. Ne on varustettu neuraalisuorittimella eli NPU:lla (neural processing unit), joka tehostaa tekoälyyn perustuvaa laskentaa.
Copilot+-kannettavien yleistymistä hidastavat ennen kaikkea korkea hinta ja ohjelmistoyhteensopivuusongelmat. Niiden keskihinta on 1 120 euroa, mikä on huomattavasti enemmän kuin perinteisten kannettavien 712 euron keskihinta. Lisäksi ARM-pohjaisen Qualcommin Snapdragon X -suorittimen käyttö on aiheuttanut yhteensopivuusongelmia joidenkin Windows-sovellusten kanssa.
Analyytikot uskovat, että tekoäly-PC:iden kysyntä kasvaa ajan myötä, mutta kehitys on hitaampaa kuin alalla alun perin ennakoitiin.
Tomi Engdahl says:
Retrieval Augmented Generation as a Service (RaaS)
https://medium.com/@asunsada/retrieval-augmented-generation-as-a-service-raas-444c797a6a27
If you ask ChatGPT 3.5 about the latest earnings results from Google or any other public company, or you ask to compare the AI strategy for different companies, ChatGPT will not be able to do so and will recommend you to check official press releases and other documents.
OpenAI’s LLM, the foundational model for ChatGPT, along with other LLMs, are dependent on the data or the domain knowledge that was used to train them (a mixture of licensed data, data created by human trainers, and publicly available data) and don’t possess real-time training capabilities. LLMs, while powerful, face challenges in independently verifying information or disclosing specific data sources.
Fortunately, with a technique called Retrieval Augmented Generation (RAG), enterprises can seize the opportunity to surpass these limitations and enhance their use cases. RAG is a technique to enhance prompts with domain-specific data to empower the LLM to generate responses that augment foundational models. RAG is rapidly becoming the standard framework for implementing enterprise applications powered by large language models (LLMs).
Despite RAG offering a less complex and resource-intensive enhancement for LLM responses compared to fine-tuning or other techniques, its implementation requires MLOps expertise, combining Data Engineering, ML engineering, and Application engineering. Enterprises venturing into GenAI experimentation also face decisions on security, privacy, scale, and price-performance while assessing business value.
Consuming RAG as a service (RaaS) emerges as a best practice for enterprises relieving users from navigating complexities and allowing them to focus on specific application requirements. RaaS simplifies the end user application-building process, making it scalable and reducing intricacies.
Although there are some RAG vendors, such as, are Nuclia, Codi or the OpenAI Assistant, this article focuses on implementing RAG internally and making it a key service as part of the enterprise AI platform.
Steps to implement a RAG Platform
The RAG platform involves the design and implementation of key capabilities, including domain data ingestion, retrieval, prompting, and RAG content generation, as well as the consumption of the resulting RAG content. It should also be equipped to deploy and operate at scale, with governance ensuring the right data access, privacy, security, and cost management for enterprise scenarios.
Tomi Engdahl says:
https://www.deepchecks.com/glossary/rag-as-a-service/
Tomi Engdahl says:
Arduino Artificial Intelligence
https://store.arduino.cc/en-fi/collections/artificial-intelligence?srsltid=AfmBOopPR_MWriHZfY8jECQYPtRU7o2heSh3Zp3SKbgTQbLHakKVE6Dy
Arduino Nano 33 BLE Sense Rev2 with headers
The Arduino Nano 33 BLE Sense Rev2 with headers is Arduino’s 3.3V AI enabled board in the smallest available form factor with a set of sensors that will allow you without any external hardware to s…
https://store.arduino.cc/en-fi/collections/artificial-intelligence/products/nano-33-ble-sense-rev2-with-headers
The Arduino Nano 33 BLE Sense Rev2 with headers is Arduino’s 3.3V AI enabled board in the smallest available form factor with a set of sensors that will allow you without any external hardware to start programming your next project, right away.
With the Arduino Nano 33 BLE Sense Rev2, you can:
Build wearable devices that using AI can recognize movements.
Build a room temperature monitoring device that can suggest or modify changes in the thermostat.
Build a gesture or voice recognition device using the microphone or the gesture sensor together with the AI capabilities of the board.
The main feature of this board, besides the complete selection of sensors, is the possibility of running Edge Computing applications (AI) on it using TinyML. Learn how to use the Tensor Flow Lite library following this instructions or learn how to train your board using Edge Impulse.
https://docs.arduino.cc/tutorials/nano-33-ble-sense/edge-impulse/
https://docs.arduino.cc/tutorials/nano-33-ble-sense/get-started-with-machine-learning/
TensorFlow Lite for Microcontrollers Examples
The inference examples for TensorFlow Lite for Microcontrollers are now packaged and available through the Arduino Library Manager making it possible to include and run them on Arduino in a few clicks. In this section we’ll show you how to run them. The examples are:
micro_speech – speech recognition using the onboard microphone
magic_wand – gesture recognition using the onboard IMU
person_detection – person detection using an external ArduCam camera
Once you connect your Arduino Nano 33 BLE Sense to your desktop machine with a USB cable you will be able to compile and run the following TensorFlow examples on the board by using the Arduino Create Cloud Edito
One of the first steps with an Arduino board is getting the LED to flash. Here, we’ll do it with a twist by using TensorFlow Lite Micro to recognise voice keywords. It has a simple vocabulary of “yes” and “no.” Remember this model is running locally on a microcontroller with only 256 KB of RAM, so don’t expect commercial ‘voice assistant’ level accuracy – it has no Internet connection and on the order of 2000x less local RAM available.
Alternatively you can use try the same inference examples using Arduino IDE application.
First, follow the instructions in the next section Setting up the Arduino IDE.
In the Arduino IDE, you will see the examples available via the File > Examples > Arduino_TensorFlowLite menu in the ArduinoIDE.
Next we will use ML to enable the Arduino board to recognise gestures. We’ll capture motion data from the Arduino Nano 33 BLE Sense board, import it into TensorFlow to train a model, and deploy the resulting classifier onto the board.
The idea for this tutorial was based on Charlie Gerard’s awesome Play Street Fighter with body movements using Arduino and Tensorflow.js.
Play Street Fighter with body movements using Arduino and Tensorflow.js
https://medium.com/@devdevcharlie/play-street-fighter-with-body-movements-using-arduino-and-tensorflow-js-6b0e4734e118
https://www.tensorflow.org/js
TensorFlow.js is a library for machine learning in JavaScript
Develop ML models in JavaScript, and use ML directly in the browser or in Node.js.
Run existing models
Use off-the-shelf JavaScript models or convert Python TensorFlow models to run in the browser or under Node.js
Retrain existing models
Retrain pre-existing ML models using your own data.
Develop ML with JavaScript
Build and train models directly in JavaScript using flexible and intuitive APIs.
https://forum.arduino.cc/t/which-arduino-board-is-best-for-ai-and-ml/1264767
TinyML is a specific branch of machine learning that focuses on running machine learning models on embedded devices with inherent limitations like model size or computational complexity.
There is nothing wrong with learning for a targeted / limited scope but you will be constrained and not be able to run the “full thing” (not even talking LLM)
if you want to do that on Arduino, have a look at
https://docs.arduino.cc/tutorials/nano-33-ble-sense/get-started-with-machine-learning/
Tomi Engdahl says:
Although the tutorial author doesn’t discuss this, if you run the program from different starting points, you get different but equivalent solutions, illustrating the “black box” nature of the derived weight arrays.
This is a well know and fundamental problem with LLMs (large language models) as well, although some inroads into understanding how the weight network actually works is slowly being made.
Arduino Neural Network
An artificial neural network developed on an Arduino Uno. Includes tutorial and source code.
http://robotics.hobbizine.com/arduinoann.html
This article presents an artificial neural network developed for an Arduino Uno microcontroller board. The network described here is a feed-forward backpropagation network, which is perhaps the most common type. It is considered a good, general purpose network for either supervised or unsupervised learning. The code for the project is provided as an Arduino sketch. It is plug and play – you can upload it to an Uno and run it, and there is a section of configuration information that can be used to quickly build and train a customized network. The write-up provided here gives an overview of artificial neural networks, details of the sketch, and an introduction to some of the basic concepts employed in feed forward networks and the backpropagation algorithm.
The sketch is available for download by clicking here: ArduinoANN.zip. The code is also listed in its entirety at the end of the tutorial.
Backpropagation neural networks have been in use since the mid-1980s. The basic concepts of backpropagation are fairly straightforward and while the algorithm itself involves some higher order mathematics, it is not necessary to fully understand how the equations were derived in order to apply them. There are some challenges to implementing a network on a very small system, and on earlier generations of inexpensive microcontrollers and hobbyist boards those challenges were significant. However Arduinos, like many of today’s boards, actually make pretty short work of the task. The Arduino Uno used here is based on Atmel’s ATmega328 microcontroller. Its 2K of SRAM is adequate for a sample network with 7 inputs and 4 outputs, and with Arduino’s GCC language support for multidimensional arrays and floating point math, the job of programming is very manageable.
So what’s it good for? Neural networks learn by example. They have been used in applications that range from autonomous vehicle control, to game playing, to facial recognition, to stock market analysis. Most applications will involve some type of pattern matching where the exact input to a system won’t be known and where there may be missing or extraneous information. Consider the problem of recognizing handwritten characters. The general shapes of the alphabet can be known ahead of time, but the actual input will always vary. Of course, the little network built here on an ATmega328 won’t be quite up to the task of facial recognition, but there are quite a few experiments in robotic control and machine learning that would be within its grasp.
In a software-based artificial neural network, neurons and their connections are constructed as mathematical relationships. When the software is presented with an input pattern, it feeds this pattern through the network, systematically adding up the inputs to each neuron, calculating the output for that neuron, and using that output to feed the appropriate inputs to other neurons in the network.
Determining the strength of the connections between neurons, also known as the weights, becomes the principal preoccupation in neural network application. In the backpropagation algorithm, the network is originally initialized with random weights. The network is then presented with a training set of inputs and outputs. As the inputs are fed through the system, the actual output is compared to the desired output and the error is calculated. This error is then fed back through the network and the weights are adjusted incrementally according to a learning algorithm. Over a period of many cycles, typically thousands, the network will ultimately be trained and will give the correct output when presented with an input.
In the feed-forward network we’re building here, the neurons are arranged in three layers called the input, hidden, and output layers. All the neurons in one layer are connected to all the neurons in the next layer. The classic graphic representation of this relationship is pictured below.
http://robotics.hobbizine.com/ArduinoANN.zip
Tomi Engdahl says:
https://etn.fi/index.php/13-news/17155-deepseek-on-jo-kielletty-monessa-kaeytoessae
Yksi merkittävimmistä huolenaiheista DeepSeekin käytössä on sen tietosuojakäytäntö. Sovellus kerää laajasti käyttäjätietoja, mukaan lukien näppäinpainallusten rytmit ja käyttäytymismallit, jotka saattavat säilyä järjestelmässä pysyvästi. Lisäksi DeepSeekin kumppanit, kuten mainostajat, jakavat tietoa käyttäjien toiminnasta myös sovelluksen ulkopuolella.
DeepSeekin tietoturvaongelmat ja sensuuriepäilyt ovat johtaneet siihen, että yhä useammat maat ja organisaatiot ovat rajoittaneet sen käyttöä. Tähän mennessä DeepSeek on kielletty tai rajoitettu seuraavissa paikoissa:
Italia – täydellinen käyttökielto
Taiwan – hallituksen virastot eivät saa käyttää DeepSeekia
Yhdysvallat – kielletty Pentagonissa, Yhdysvaltain kongressissa, laivastossa, NASAssa ja Texasissa
Australia – kielletty kaikilla valtionhallinnon laitteilla
AIPRM:n perustaja Christoph C. Cemperin mukaan DeepSeekin tietosuojakäytännöt ovat herättäneet erityistä huolta. – Tämä on yksi syy siihen, miksi useat maat ovat jo ryhtyneet rajoittamaan sen käyttöä. Yhdysvalloissa on jopa esitetty lakiehdotusta, jonka mukaan DeepSeekin käyttö voisi johtaa sakkoihin tai vankeusrangaistukseen.
Tutkimusten mukaan DeepSeekin turvatoimet ovat helposti ohitettavissa, mikä mahdollistaa haitallisen sisällön, kuten vihapuheen, uhkaukset ja rikolliseen toimintaan liittyvän materiaalin leviämisen sovelluksen kautta. Lisäksi tutkimuksissa havaittiin, että peräti 83 % sovelluksen testeissä tuotetuista vastauksista sisälsi syrjiviä elementtejä.
https://www.aiprm.com/
Tomi Engdahl says:
A robot designed for war recently performed a 30-minute DJ set at a nightclub in San Francisco. No, seriously, this is a thing that happened. A reporter was there and saw it.
Culture
A Robot Designed For War Did A 30-Minute DJ Set At A California Club
https://brobible.com/culture/article/robot-war-dj-california-club/?fbclid=IwY2xjawIgJ_FleHRuA2FlbQIxMQABHUIDjo4cuJuCrQJAHrXSspwD0rHIOUCyKsgAs9X-kcEztVvY_C86ZohSvg_aem_lqlfCV6PpqnquetlylsOYA
A robot designed for war recently performed a 30-minute DJ set at a nightclub in San Francisco. No, seriously, this is a thing that happened. A reporter was there and saw it.
Timothy Karoff, Culture Reporter for SFGate.com, recently went to Temple Nightclub to see Phantom, the first humanoid robot developed by San Francisco-based startup Foundation Robotics Labs, rock the house. Foundation, by the way, is the only American robotics company building humanoids for the military. And laying down some dope tracks, apparently.
Tomi Engdahl says:
Kyle Wiggers / TechCrunch:
xAI launches Grok3 beta and Grok3 mini, its latest AI models with reasoning capabilities, trained on “10x” more compute than Grok2, for X Premium+ subscribers — Elon Musk’s AI company, xAI, late on Monday released its latest flagship AI model, Grok 3, and unveiled new capabilities for the Grok iOS and web apps.
Elon Musk’s xAI releases its latest flagship model, Grok 3
https://techcrunch.com/2025/02/17/elon-musks-ai-company-xai-releases-its-latest-flagship-ai-grok-3/
Tomi Engdahl says:
Amy Thomson / Bloomberg:
xAI unveils DeepSearch, a reasoning chatbot that explains its thought process for queries and is capable of doing research, brainstorming, and data analysis — Elon Musk’s artificial intelligence startup xAI showed off the updated Grok-3 model, showcasing a version of the chatbot technology …
https://www.bloomberg.com/news/articles/2025-02-18/musk-s-xai-debuts-grok-3-ai-bot-touting-benchmark-superiority
Tomi Engdahl says:
Andrej Karpathy / @karpathy:
Grok3 review: its thinking capability feels state of the art and rivals OpenAI’s o1 pro models, DeepSearch offers a blend of search and reasoning, and more
https://x.com/karpathy/status/1891720635363254772
First, Grok 3 clearly has an around state of the art thinking model (“Think” button) and did great out of the box on my Settler’s of Catan question:
“Create a board game webpage showing a hex grid, just like in the game Settlers of Catan. Each hex grid is numbered from 1..N, where N is the total number of hex tiles. Make it generic, so one can change the number of “rings” using a slider. For example in Catan the radius is 3 hexes. Single html page please.”
Few models get this right reliably. The top OpenAI thinking models (e.g. o1-pro, at $200/month) get it too, but all of DeepSeek-R1, Gemini 2.0 Flash Thinking, and Claude do not.
I uploaded GPT-2 paper. I asked a bunch of simple lookup questions, all worked great. Then asked to estimate the number of training flops it took to train GPT-2, with no searching.
The impression overall I got here is that this is somewhere around o1-pro capability, and ahead of DeepSeek-R1, though of course we need actual, real evaluations to look at.
DeepSearch
Very neat offering that seems to combine something along the lines of what OpenAI / Perplexity call “Deep Research”, together with thinking. Except instead of “Deep Research” it is “Deep Search” (sigh). Can produce high quality responses to various researchy / lookupy questions you could imagine have answers in article on the internet, e.g. a few I tried, which I stole from my recent search history on Perplexity, along with how it went
Summary. As far as a quick vibe check over ~2 hours this morning, Grok 3 + Thinking feels somewhere around the state of the art territory of OpenAI’s strongest models (o1-pro, $200/month), and slightly better than DeepSeek-R1 and Gemini 2.0 Flash Thinking. Which is quite incredible considering that the team started from scratch ~1 year ago, this timescale to state of the art territory is unprecedented. Do also keep in mind the caveats – the models are stochastic and may give slightly different answers each time, and it is very early, so we’ll have to wait for a lot more evaluations over a period of the next few days/weeks. The early LM arena results look quite encouraging indeed. For now, big congrats to the xAI team, they clearly have huge velocity and momentum and I am excited to add Grok 3 to my “LLM council” and hear what it thinks going forward.
Tomi Engdahl says:
William Boston / Wall Street Journal:
How militaries and startups like North.io use AI to analyze data and deploy autonomous underwater vehicles to safeguard deep-sea pipelines and cables
How AI Can Protect Vital Pipelines and Cables Deep in the Ocean
Militaries and startups use artificial intelligence to sift through vast amounts of data and power autonomous underwater vehicles, boosting efforts to surveil the seabed
https://www.wsj.com/tech/ai/ai-military-applications-mapping-aca7f486?st=dUSNMX&reflink=desktopwebshare_permalink
Deep under the sea, pipelines and cables carrying fuel, power and communications are strewn on the ocean floor like a central nervous system for the global economy.
Huge stretches of these critical connectors lie unprotected in the murky depths—and vulnerable to attacks such as the 2022 sabotage of the Nord Stream pipelines that carry Russian natural gas to Europe under the Baltic Sea.
Now, in the way that the use of drones has changed the conduct of land wars, artificial intelligence is about to change everything about how the deep sea is navigated and how critical underwater infrastructure is protected in wartime and against threats of terrorism.
It’s hard to get a look at this undersea world in order to protect it. Data is fragmented and comes from systems ranging from sonar to satellites. Analyzing it often takes weeks. What’s needed is a sort of Google Maps of the sea that not only accurately replicates the oceans and their terrain but also provides timely alerts on potential threats. That’s where AI’s ability to sift through vast amounts of data comes in.
AI-empowered underwater systems are already changing seabed warfare and defense. Drones and anti-mine robots, working together with ships on the surface, underwater sensors and satellites, are being deployed by the military and governments. These systems to navigate, map, and provide underwater defense are increasingly using AI to analyze and synthesize diverse sources of data.
In the next few years, industry and military users expect these systems to make huge advances.
Tomi Engdahl says:
Romain Dillet / TechCrunch:
Mistral debuts Mistral Saba, a 24B-parameter custom-trained model for Arabic language and culture, via its API; Saba bests Mistral Small 3 for Arabic content
Mistral releases regional model focused on Arabic language and culture
https://techcrunch.com/2025/02/17/mistral-releases-regional-model-focused-on-arabic-language-and-culture/
The next frontier for large language models (LLMs), one of the key technologies underpinning the boom in generative AI tools, might be geographical. On Monday, Paris-based AI startup Mistral — which is vying to rival the likes of U.S.-based Anthropic and OpenAI — is releasing a model that’s a bit different from its usual LLM.
Named Mistral Saba, the new custom-trained model is designed to address a specific geography: Arabic speaking countries. The goal for Mistral Saba is to excel in Arabic interactions.
Tomi Engdahl says:
https://hackaday.com/2025/02/13/why-ai-usage-may-degrade-human-cognition-and-blunt-critical-thinking-skills/
Tomi Engdahl says:
https://hackaday.com/2025/02/17/usb-stick-hides-large-language-model/
Tomi Engdahl says:
Elon Musk’s artificial intelligence company, xAI, has officially released the Grok 3 series of models. https://link.ie.social/epAZtS
Tomi Engdahl says:
Investigation on identify the multiple issues in IoT devices using Convolutional Neural Network
https://www.sciencedirect.com/science/article/pii/S266591742200143X
Current Challenges in IoT Security and Forensics: Strategies for a Secure Connected Future
https://www.intechopen.com/online-first/1207070
Tomi Engdahl says:
Train AI To Control Micro:Bit Robot With Hand Gestures
https://www.youtube.com/watch?v=RmbFNf6yOME
https://cardboard.lofirobot.com/control-robot-with-hand-gestures/
Train AI to Control Micro:Bit Robot with Hand Gestures
Introduction
Controlling robots with hand gestures is an exciting intersection of artificial intelligence (AI) and hardware. In this tutorial, you will learn how to train an AI model to control a Micro:Bit robot using hand gestures, leveraging tools like Teachable Machine and Micro:Bit’s Bluetooth connectivity.
To train an AI model, you will use an online tool called Teachable Machine. This app allows you to create machine learning models without coding. You can access it through your browser and follow these steps:
Open Teachable Machine: Go to the Teachable Machine website.
Create New Project: Choose an image project and create classes for each hand gesture, such as left, right, up, down, open, and close.
Capture Gestures: Use your webcam or smartphone camera to capture images for each gesture.
Train the Model: Click the “Train Model” button and let the app process your images.
Export the Model: Once trained, export your model and save the sharable link.
Integrating with Micro:Bit
On the Micro:Bit side, the board receives class names via Bluetooth. Here’s how you can integrate the AI model with your robot:
Open MakeCode: Go to the MakeCode editor and create a new project.
Code the Movements: Use simple if conditions to map each class name to specific motor actions, such as turning left or right.
Enable Bluetooth: Ensure that your project settings enable Bluetooth discoverability – set it to JUST WORKS
https://cardboard.lofirobot.com/teachable-microbit-app-info/
Teachable Micro:Bit is a progressive web app (compatible with desktop and mobile devices) that allows you to connect a machine learning model trained in Teachable Machine with a Micro:Bit board, adding basic AI capabilities to your robots.
Teachable Machine is an educational web app developed by Google that simplifies the process of training a machine learning model for image, sound, and pose classification tasks. Using a computer webcam, you can easily upload photos of your training data and, within minutes, have a functioning model capable of recognizing hand gestures, objects, patterns, and more. This makes Teachable Machine a fantastic educational tool in the classroom for demonstrating the principles of machine learning and helping young students grasp the basics of Artificial Intelligence in practice.
Tomi Engdahl says:
..uudet GPU:t tuntuu olevan kyllä pettymyksiä. Kaikki effortti lienee laitettu AI:hin, graffasuorituskyvyn parannukset on jäämässä marginaalisiksi.
Tomi Engdahl says:
AI Genius – Session 4 – Building Intelligent Multi-agent Systems
https://www.youtube.com/watch?v=IWXWTigK0Ic
Explore how Azure AI Foundry seamlessly integrates with agents to create intelligent, responsive systems. We’ll open up the session with Thermo Fisher who has created a multi-agent system to query across a vast store of disparate data. Marco and Mads will then walk you through the new Agent Service, offering practical examples and best practices for implementing AI agents in various business contexts.
Tomi Engdahl says:
Meet the AI taking your job
https://www.youtube.com/shorts/7Zcel8PFKNk
Tomi Engdahl says:
Large Language Models explained briefly
https://www.youtube.com/watch?v=LPZh9BOjkQs