Tools & Platforms
Thinking Machines Lab wants to make AI models more consistent

There’s been great interest in what Mira Murati’s Thinking Machines Lab is building with its $2 billion in seed funding and the all-star team of former OpenAI researchers who have joined the lab. In a blog post published on Wednesday, Murati’s research lab gave the world its first look into one of its projects: creating AI models with reproducible responses.
The research blog post, titled “Defeating Nondeterminism in LLM Inference,” tries to unpack the root cause of what introduces randomness in AI model responses. For example, ask ChatGPT the same question a few times over, and you’re likely to get a wide range of answers. This has largely been accepted in the AI community as a fact — today’s AI models are considered to be non-deterministic systems— but Thinking Machines Lab sees this as a solvable problem.
The post, authored by Thinking Machines Lab researcher Horace He, argues that the root cause of AI models’ randomness is the way GPU kernels — the small programs that run inside of Nvidia’s computer chips — are stitched together in inference processing (everything that happens after you press enter in ChatGPT). He suggests that by carefully controlling this layer of orchestration, it’s possible to make AI models more deterministic.
Beyond creating more reliable responses for enterprises and scientists, He notes that getting AI models to generate reproducible responses could also improve reinforcement learning (RL) training. RL is the process of rewarding AI models for correct answers, but if the answers are all slightly different, then the data gets a bit noisy. Creating more consistent AI model responses could make the whole RL process “smoother,” according to He. Thinking Machines Lab has told investors that it plans to use RL to customize AI models for businesses, The Information previously reported.
Murati, OpenAI’s former chief technology officer, said in July that Thinking Machines Lab’s first product will be unveiled in the coming months, and that it will be “useful for researchers and startups developing custom models.” It’s still unclear what that product is, or whether it will use techniques from this research to generate more reproducible responses.
Thinking Machines Lab has also said that it plans to frequently publish blog posts, code, and other information about its research in an effort to “benefit the public, but also improve our own research culture.” This post, the first in the company’s new blog series called “Connectionism,” seems to be part of that effort. OpenAI also made a commitment to open research when it was founded, but the company has become more closed off as it’s become larger. We’ll see if Murati’s research lab stays true to this claim.
The research blog offers a rare glimpse inside one of Silicon Valley’s most secretive AI startups. While it doesn’t exactly reveal where the technology is going, it indicates that Thinking Machines Lab is tackling some of the largest question on the frontier of AI research. The real test is whether Thinking Machines Lab can solve these problems, and make products around its research to justify its $12 billion valuation.
Techcrunch event
San Francisco
|
October 27-29, 2025
Tools & Platforms
5-Week AI Mentorship for Startups in SF

OpenAI has unveiled a new initiative aimed at nurturing the next generation of artificial intelligence innovators, marking a strategic push into talent development amid intensifying competition in the AI sector. The program, dubbed OpenAI Grove, targets early-stage entrepreneurs who are either pre-idea or in the nascent phases of building AI-focused companies. According to details shared in a recent announcement, the five-week mentorship scheme will be hosted at OpenAI’s San Francisco headquarters, providing participants with hands-on guidance from industry experts and access to cutting-edge tools.
The program’s structure emphasizes practical support, including technical assistance, community building, and early exposure to unreleased OpenAI models. As reported by The Indian Express, participants will have opportunities to interact with new AI tools before their public release, fostering an environment where budding founders can experiment and iterate rapidly. This comes at a time when AI startups are proliferating, with OpenAI positioning itself as a hub for innovation rather than just a technology provider.
A Strategic Move in AI Talent Cultivation OpenAI’s launch of Grove reflects a broader effort to secure its influence in the rapidly evolving AI ecosystem, where retaining and attracting top talent is crucial. By offering mentorship to pre-seed founders, the company aims to create a pipeline of AI-driven ventures that could potentially integrate with or complement its own technologies. Recent posts on X highlight enthusiasm from the tech community, with users noting the program’s potential to accelerate startup growth through exclusive access to OpenAI’s resources.
Industry observers see this as OpenAI’s response to competitors like Anthropic and Grok, which have also been aggressive in talent acquisition. The first cohort, limited to about 15 participants, is set to run from October 20 to November 21, 2025, with applications closing on September 24. As detailed in coverage from CNBC, the initiative includes in-person sessions focused on co-building prototypes with OpenAI researchers, underscoring a hands-on approach that differentiates it from traditional accelerator programs.
Benefits and Broader Implications for Startups Participants in Grove stand to gain more than just technical know-how; the program promises a robust network of peers and mentors, which could be invaluable for fundraising and scaling. Early access to unreleased models, as mentioned in reports from NewsBytes, allows founders to test ideas with state-of-the-art AI capabilities, potentially giving them a competitive edge in a market where speed to innovation is key.
This mentorship model aligns with OpenAI’s history of fostering external ecosystems, similar to its past investments in startups through funds like the OpenAI Startup Fund. However, Grove appears more focused on individual founders, particularly those without formal teams or funding, addressing a gap in the startup support system. Insights from The Daily Jagran emphasize how the program could help participants raise capital or refine their business models, drawing on expert guidance to navigate challenges like ethical AI development and market fit.
Challenges and Future Outlook While the program has generated buzz, questions remain about its scalability and inclusivity. With only 15 spots in the initial cohort, selection will be highly competitive, potentially favoring founders with existing connections in the tech world. Recent news on X suggests mixed sentiments, with some praising the initiative for democratizing AI access, while others worry it might reinforce Silicon Valley’s dominance in the field.
Looking ahead, OpenAI plans to run Grove multiple times a year, potentially expanding its reach globally. As covered in TechStory, this could evolve into a cornerstone of OpenAI’s strategy to build a supportive community around its technologies, much like how Y Combinator has shaped the broader startup world. For industry insiders, Grove represents not just a mentorship opportunity but a signal of OpenAI’s commitment to shaping the future of AI entrepreneurship, ensuring that innovative ideas flourish under its umbrella.
Potential Impact on the AI Innovation Ecosystem The introduction of Grove could catalyze a wave of AI startups, particularly in areas like generative models and ethical AI applications, by providing resources that lower barriers to entry. Founders selected for the program will benefit from personalized feedback loops, helping them avoid common pitfalls in AI development such as data biases or scalability issues.
Moreover, this initiative underscores OpenAI’s evolution from a research lab to a multifaceted player in the tech industry. By mentoring early-stage talent, the company may indirectly fuel advancements that enhance its own ecosystem, creating a virtuous cycle of innovation. As the AI sector continues to mature, programs like Grove could play a pivotal role in distributing expertise more evenly, empowering a diverse array of entrepreneurs to contribute to technological progress.
Tools & Platforms
San Antonio Spa Unveils First AI-Powered Robot Massager

In the heart of San Antonio, a quiet revolution in wellness technology is unfolding at Float Wellness Spa on Fredericksburg Road. The spa has become the first in the city to introduce the Aescape AI-powered robot massager, a device that promises to blend cutting-edge artificial intelligence with the ancient art of massage therapy. Customers lie face-down on a specialized table, where robotic arms equipped with sensors scan their bodies to deliver personalized treatments, adjusting pressure and techniques in real time based on individual anatomy and preferences.
This innovation arrives amid a broader surge in AI applications within the health and wellness sector, where automation is increasingly tackling labor shortages and consistency issues in human-delivered services. According to a recent feature by Texas Public Radio, the Aescape system at Float Wellness Spa uses advanced algorithms to map muscle tension and provide targeted relief, marking a significant step for Texas in adopting such tech.
Technological Backbone and Operational Mechanics
At its core, the Aescape robot employs a combination of 3D body scanning, machine learning, and haptic feedback to simulate professional massage techniques. Users select from various programs via a touchscreen interface, and the system adapts on the fly, much like a therapist responding to subtle cues. This isn’t mere gimmickry; it’s backed by years of development, with the company raising substantial funds to refine its precision.
In a March 2025 report from Bloomberg, Aescape secured $83 million in funding from investors including Valor Equity Partners and NBA star Kevin Love, underscoring investor confidence in robotic wellness solutions. The technology draws from earlier prototypes showcased at events like CES 2024, where similar AI-driven massage robots demonstrated personalized adaptations to user needs.
Market Expansion and Local Adoption in San Antonio
The rollout in San Antonio follows successful debuts in cities like Los Angeles, as detailed in a December 2024 piece by the Los Angeles Times, which described the experience as precise yet impersonal. At Float Wellness Spa, appointments are now bookable, with sessions priced competitively to attract a mix of tech enthusiasts and those seeking convenient relief from daily stresses.
Posts on X, formerly Twitter, reflect growing public intrigue, with users like tech influencer Mario Nawfal highlighting the robot’s eight axes of motion for deep-tissue work without the awkwardness of human interaction. This sentiment aligns with San Antonio’s burgeoning tech scene, where AI innovations are intersecting with local industries, as noted in recent updates from the San Antonio Express-News.
User Experiences and Industry Implications
Early adopters in San Antonio report a mix of awe and adjustment. One reviewer in a Popular Science article from March 2024 praised the Aescape for its customized convenience, likening it to “the world’s most advanced massage” powered by AI that learns from each session. However, some note the absence of human warmth, a point echoed in an Audacy video report from August 2025, which captured the robot’s debut turning heads in the city.
For industry insiders, this represents a pivot toward scalable wellness tech. With labor costs rising and therapist shortages persistent, robots like Aescape could redefine spa economics, potentially expanding to chains like Equinox. Yet, challenges remain, including regulatory hurdles for AI in healthcare-adjacent fields and ensuring data privacy for body scans.
Future Prospects and Competitive Dynamics
Looking ahead, Aescape’s expansion signals broader trends in robotic automation. A Yahoo Finance piece from August 2025 introduced a competing system, RoboSculptor, which also leverages AI for massage, hinting at an emerging market rivalry. In San Antonio, this could spur further innovation, with local startups like those covered in Nucamp’s tech news roundup exploring AI tools in customer service and beyond.
As AI integrates deeper into personal care, ethical questions arise—will robots supplant human jobs, or augment them? For now, Float Wellness Spa’s offering provides a tangible glimpse into this future, blending Silicon Valley ingenuity with Texas hospitality. Industry watchers will be keen to monitor adoption rates, as success here could accelerate nationwide rollout, transforming how we unwind in an increasingly automated world.
Tools & Platforms
California AI Regulation Bill SB 1047 Stalls Amid Tech Lobby Pushback

California’s ambitious push to regulate artificial intelligence has hit another snag, with key legislation stalled amid fierce debates over innovation, safety, and economic impact. Lawmakers had high hopes for 2025, building on previous efforts like the vetoed SB 1047, but recent developments suggest a familiar pattern of delay. According to a report from CalMatters, the state’s proposed AI safety bill, SB 53, which aimed to impose strict testing and oversight on advanced models, remains in limbo as Governor Gavin Newsom weighs his options. This comes after a year of intense lobbying from tech giants and startups alike, highlighting the tension between fostering cutting-edge tech and mitigating potential risks.
The bill’s provisions, including mandatory safety protocols for models trained with massive computational power, have drawn both praise and criticism. Proponents argue it could prevent catastrophic misuse, such as AI-driven cyberattacks or misinformation campaigns, while opponents warn it might stifle California’s tech dominance. Newsom’s previous veto of similar measures cited concerns over overregulation, a sentiment echoed in recent industry feedback.
The Political Tug-of-War Intensifying in Sacramento
As the legislative session nears its end, insiders point to behind-the-scenes negotiations that have bogged down progress. Sources from White & Case LLP note that while some AI bills, like the Generative AI Accountability Act, were signed into law effective January 1, 2025, broader safety frameworks face resistance. This act requires state agencies to conduct risk analyses and ensure ethical AI use, but it stops short of comprehensive private-sector mandates. Meanwhile, posts on X from tech figures like Palmer Luckey express relief over potential federal pre-emption, suggesting that national guidelines might override state efforts to avoid a patchwork of rules.
The delay’s roots trace back to economic pressures. California’s tech sector, home to Silicon Valley heavyweights, contributes massively to the state’s GDP. A Inside Global Tech analysis reveals that over a dozen AI bills advanced this session, covering consumer protections and chatbot safeguards, yet core safety bills like SB 53 are caught in crossfire. Industry leaders argue that vague liability clauses could drive companies to relocate, with estimates from X discussions indicating potential job losses in the thousands.
Economic Ramifications and Industry Pushback
Compliance costs are a flashpoint. A study referenced in posts on X by Will Rinehart, using large language models to model expenses, projects that firms could face $2 million to $6 million in burdens over a decade for automated decision systems under bills like AB 1018. This has mobilized opposition from companies like Anthropic, which paradoxically endorsed some regulations but lobbied against overly burdensome ones, as reported by NBC News via X updates. Startups, in particular, fear being crushed under regulatory weight that Big Tech can absorb, with TechCrunch highlighting how SB 243’s chatbot rules could set precedents for accountability without derailing innovation.
Governor Newsom’s decision looms large, influenced by his national ambitions and the state’s budget woes. Recent web searches show a June 2025 expert report, The California Report on Frontier AI Policy, informing revisions to make the bill less “rigid,” per Al Mayadeen English. Yet, delays persist, with critics on X platforms like @amuse warning that California risks ceding AI leadership to China if regulations become too stringent.
Looking Ahead: Innovation vs. Safeguards
The holdup underscores a broader national debate. While California has enacted laws on deepfakes and AI transparency—such as AB 2013 requiring training data disclosure, as detailed by Mayer Brown—comprehensive AI governance remains elusive. Experts predict that without resolution by year’s end, federal intervention could preempt state actions, a scenario favored by some X commentators like Just Loki.
For industry insiders, this delay offers a reprieve but also uncertainty. Companies are already adapting, with some shifting operations to states like Texas for lighter oversight. As Pillsbury Law outlines, the 18 new AI laws effective in 2025 focus on sectors like healthcare and elections, yet the absence of overarching safety nets leaves gaps. Ultimately, California’s AI regulatory saga reflects the high stakes: balancing technological progress with societal protection in an era where AI’s potential—and perils—are only beginning to unfold.
-
Business2 weeks ago
The Guardian view on Trump and the Fed: independence is no substitute for accountability | Editorial
-
Tools & Platforms1 month ago
Building Trust in Military AI Starts with Opening the Black Box – War on the Rocks
-
Ethics & Policy2 months ago
SDAIA Supports Saudi Arabia’s Leadership in Shaping Global AI Ethics, Policy, and Research – وكالة الأنباء السعودية
-
Events & Conferences4 months ago
Journey to 1000 models: Scaling Instagram’s recommendation system
-
Jobs & Careers2 months ago
Mumbai-based Perplexity Alternative Has 60k+ Users Without Funding
-
Podcasts & Talks2 months ago
Happy 4th of July! 🎆 Made with Veo 3 in Gemini
-
Education2 months ago
Macron says UK and France have duty to tackle illegal migration ‘with humanity, solidarity and firmness’ – UK politics live | Politics
-
Education2 months ago
VEX Robotics launches AI-powered classroom robotics system
-
Funding & Business2 months ago
Kayak and Expedia race to build AI travel agents that turn social posts into itineraries
-
Podcasts & Talks2 months ago
OpenAI 🤝 @teamganassi