Beyond the Mainframe Nº5 | Laika Ventures
AI giants reshuffle ♟️ New global alliances 🌍 Building has never been easier
Welcome to Laika Ventures’ monthly newsletter! We promise to bring you the most relevant news on AI, fire 🔥 discussions, and more.
We all know trying to stay updated on what’s happening in AI can be difficult with so much going on and how fast it is evolving, so that’s why we cover last month’s relevant AI news.
In May 2025, we saw OpenAI walk back its for-profit pivot, new alliances forming between tech giants and nations, conversations on regulating AI took place, and a rise in agents as the new coding coworkers. Startups and countries alike are still racing to define their place in the AI-powered world: from national strategies like the UAE’s ChatGPT rollout to AI-native browsers and creative tools like Figma Make and Veo 3 — May proved that the AI race isn’t slowing down, it’s leveling up. 🚀
Let’s go!
Open AI’s news updates 🗞️
You probably remember the scandal November 17 2023 that got everyone perplexed for a couple of days when Sam Altman was fired from Open AI. We were able to understand the structural mess in that company.
Earlier this month, OpenAI announced the backing-out of their for-profit ambitions. After months of lawsuits and protests, the non-profit will retain control of OpenAI. “The for-profit LLC under the non-profit will transition to a Public Benefit Corporation (PBC) with the same mission”.
OpenAI just announced a new initiative within the Stargate project (the $500B investment in America’s AI infrastructure) to support countries around the world to “build on democratic AI rails, and provide a clear alternative to authoritarian versions of AI that would deploy it to consolidate power”, starting with UAE 🇦🇪.
Earlier this month, OpenAI also announced the most recent executive hire - Fidji Simo, current CEO of 🥕 Instacart - to join as Chief of Applications, and to help the company grow to its new phase, from a research lab to a multi-faceted global organization.
One of the most discussed topics relating to OpenAI was the acquisition of Jony Ive’s AI device startup, io, in a $6.5B all-stock deal. Ive remains independent but becomes OpenAI’s de-facto chief creative officer, and LoveFrom studio will “assume deep design responsibilities” across every OpenAI product. This marks the intention of OpenAI to control the last mile to the user without depending on companies like Apple, Google or Microsoft.
Regulation is still being discussed 📄
The USA 🇺🇸 Senate is examining how the country can maintain and strengthen its global leaderhsip in AI. On 8th of May, the hearing - titled “Winning the AI Race: Strengthening U.S. Capabilities in Computing and Innovation” - was convened to discuss how removing regulatory barriers could accelerate development, production and availability of AI technologies. Major tech executives gave testimony to the hearing:
Mr. Sam Altman, Co-Founder and Chief Executive Officer, OpenAI
Dr. Lisa Su, Chief Executive Officer and Chair, Advanced Micro Devices (AMD)
Mr. Michael Intrator, Co-Founder and Chief Executive Officer, CoreWeave
Mr. Brad Smith, Vice Chair and President, Microsoft Corporation
Brad Smith mentioned, during the hearing, “At Microsoft we don’t allow our employees to use the DeepSeek app. We did not put the DeepSeek app in our app store”. This was the first time that Microsoft has gone public about this ban.
The way to beat China in the AI race is to outrace them in innovation, not saddle AI developers with European-style regulations. Growth and development of new AI technologies will bolster our national security, create new jobs, and stimulate economic growth.
- U.S. Senator Ted Cruz (R-Texas)
This month, the Trump Administration decided to reject the Biden Administration’s AI Diffusion Rule that would have imposed controls on semiconductor export worldwide, just days before the rule was set to take effect. At the same time, they announced to strengthen export controls for overseas AI chips including Huawei Ascend chips, which quickly caught China’s attention, urging to cease discriminatory measures.
Search world 🔎
Aravind Srinivas, CEO of Perplexity, tweeted about the new feature of the company - Perplexity Labs - that can do complex tasks, like pulling financial reports.
Alibaba introduced ZeroSearch, a technique that teaches AI systems to search without using real search engines, which cuts training costs and can perform better than models who use search APIs. While on the topic of news ways to train AI models, researchers from Tsinghua Univerisity and BIGAI (Beijing Institute for General Artificial Intelligence) introduced a technique where models learn and master complex tasks on their own through self-play - it’s called “Absolute Zero”.
The Browser Company announced its strategy shift moving away from the Arc browser to an AI browser called Dia.
Coding models race 🏁
Mistral released a cost-effective AI model, even though its performance is still a step down from top models. They’ve released Le Chat Enterprise for those with strict security requirements, and those concern with data governance under EU regulations. Later this month, Mistral and All Hands AI introduced an open-source model for coding agents that would outperform other competitors while being small enough to run on a laptop or single GPU.
The very next day after Mistral’s announcement, Anthropic released what they would call the “world’s best coding model” - Claude 4.
Microsoft introduced a coding agent for GitHub Copilot. The agent is embedded into GitHub and starts to work when you assign a GitHub issue to Copilot.
This month’s new releases on coding agent reminds us of the shift that’s happening in how software is being built - instead of writing code, a developer can delegate tasks to agents and focus on strategy and problem-solving. Building became much faster.
Figma is ready for the fight 🥊
During the Config 2025 event, Figma made sure its positioning in AI doesn’t go unnoticed. Figma announced 4 new products that compete with other giants (Canva, Adobe, WebFlow, Framer).
Figma Make - a promp-to-code to transform designs into interactive prototypes.
Figma Sites - an all-in-one tool to design and build custom websites.
Figma Draw - faster, simpler vector editing with powerful tools for visual expression.
Figma Buzz - shared space for designers and marketers to work together.
KSA and UAE making a statement 📝
UAE 🇦🇪 announced that AI would be mandatory for all K-12 students starting this year.
In another effort to make sure their people are prepared for the AI age, UAE 🇦🇪 “will become the first country in the world to enable ChatGPT nationwide” by acquiring ChatGPT Plus ($20) subscriptions for the its entire population. This is a huge step in public AI access.
Meanwhile, in the Kingdom of Saudi Arabia 🇸🇦, his Royal Highness Prince Mohammed bin Salman bin Abdulaziz Al Saud, Crown Prince, Prime Minister and Chairman of the Board of Directors of Public Investment Fund (PIF), announced the launch of HUMAIN, a PIF-owned company. HUMAIN will operate and invest across the artificial intelligence (AI) value chain as a unified operating company.
Google’s and Microsoft’s new updates 💡
While we started the year with the idea that “2025 is the year of AI agents”, it hasn’t been as practical as one would expect. During Build 2025, Satya Nadella shared Microsoft’s vision for that, and new AI tools and updates were also made. One of these new tools is Microsoft Discovery, an agentic platform to accelerate research and development (R&D).
This month Google also held the I/O event. Sundar Pichai stated “Normally, you wouldn’t have heard much from us in the weeks leading up to I/O. That’s because we’d be saving our best models for this stage. But in our Gemini era, we are just as likely to ship our most intelligent model on a random Tuesday in March or a really cool breakthrough like AlphaEvolve just a week before. We want to get our best models into your hands and our products ASAP. And so we’re shipping faster than ever”.
They announced new media models for video and image generation, with Veo 3 and Imagen 4, and a filmmaker tool called Flow.
Vibecoding is here to stay with the new release of Gemini 2.5 Pro Preview (I/O edition) with even better coding performance.
Google also created the AI Futures Fund that will provide early access to Google DeepMind’s latest AI models and other resources. Apply here.
Some more new releases 📰
NVIDIA is not only dominating the chip game, but it is also releasing powerful open-source models. The latest automatic speech recognition (ASR) model, Parakeet V2, can transcribe an hour of audio in a single second and took the top spot on the Open ASR leaderboard with 6.05% Word Error Rate.
LTX Studio launched a new open-source video model with improved speed, quality and more cost-effective.
The New York Times announced a partnership with Amazon for an AI licensing deal, while having ongoing lawsuits against OpenAI and Microsoft over the use of its contente for training.
Funny Updates 😆
Sergey Brin, Google co-founder, shared that Gemini and other models “tend to do better if you threaten them, like with physical violence”.
Image generated using GPT-4
Can regulation diminish the risks associated with AI?
Most regulation is prescriptive about what you must guarantee—data anonymization, resilience against attacks, traceability—but not how to implement it. You must think from day one about where data is stored, how it’s transformed, and how to stop/interrupt bias and prevent leaks.
There are several risks: deepfakes, synthetic-data loops, and models “optimized for engagement”, trying to imitate the human behavior, can amplify misinformation. And I don’t think we’re ready to identify synthetic and fake information, and many times people don’t even care.
There’s a big risk that we’re training models with data generated from those very own models that would diminish their quality and be used to perpetuate biases present in the data used to train these models. The saying “garbage in garbage out” used for training is also true when people are interacting with it, and testing in the real world.
- João Galego, Head of AI at Critical Software
Read the full article in our blog
We’re a portuguese company based in Braga, and we’re all about celebrating 🇵🇹 portuguese innovations!
Portugal has one more unicorn 🦄 - TEKEVER, the portuguese defense tech startup specialized in AI-powered autonomous systems.
Bloq.it, the portuguese smart locker company, raised €28M in a series B round 👏🏻
Lisbon AI Conference 2025 happened in May 22 and counted with names like Helena Moniz, Chair of the Ethics Committee of the Center for Responsible AI.
IN Conference INEGI 2025 was focused in AI and held more than 200 people at the cruise Terminal of the Port of Leixões.
Also in Porto, SIM Conference was this month’s stage for startup’s showcase, panel discussions and networking.
→ If you’re doing amazing things in Portugal and we didn’t highlight it, let us know! We keep our eyes wide open but sometimes we can miss something 😉
🐾 Join us!
Last month we were busy expanding our events! We held an event in Riyadh 🇸🇦 at the Portuguese Ambassador’s house and held our first event in Lisbon 🇵🇹 in partnership with Critical Software.
We had such a good reception in Lisbon that we will be coming back next month. Save the date 17th of June for an 💫 Orbit event at the AIhub. Sign up for the event here and for the coworking day here.
Check what we’re doing at laikaventures.co. Thanks for reading!