Latest AI News
Sesame, the conversational AI startup from Oculus founders, launches its iOS app
On Thursday, the AI startupSesame, co-founded byOculus’s foundersand others from theVR company that sold to Meta, released a public preview of the conversational AI agents it’s been developing for over a year. With itsnew iOS app, Sesame is rethinking the traditional AI chatbot experience popularized by apps like ChatGPT, creating one where conversation flows, even if the AI needs time to think. As the company explains in its launchannouncement, “There’s an inherent tension between replying quickly and taking the time to compose thoughtful responses. A slower response is usually more correct, but it can also feel unnatural if it takes too long.” To address this challenge, Sesame claims to have built fast search and retrieval systems, so the AI can have up-to-date information, as well as technology that allows it to run multiple parallel searches while speaking, weaving those results into its responses as it talks. That means the AI will talk more like a human, even pivoting mid-sentence if need be, as it taps into newer information — as a human might when remembering another key fact or point they want to add. The app offers four distinct AI agents called Maya, Miles, Simone, and Charlie, each of which have their own distinct voice, personality, point of view, and memory. Maya and Miles werepreviously availablein Sesame’s Research Preview of its technology, where they were soon accessed by over a million people within the first few weeks,said Sesame investor Sequoiaat the time. (The company had then just raised its $250 million Series B from Sequoia and others, and was opening up a beta.) During the beta, Sesame learned from user feedback and rolled out features including search cards with image results for visualizing concepts, notes for capturing takeaways, a texting mode for those times when speaking aloud is not an option, and support for deep dives where you can get more in-depth results. There’s also a new incognito mode for private conversations, which allows the agents access to prior context, but saves nothing to memory. The app, however, is only the first step towards Sesame’s bigger plans for AI involving intelligent eyewear, which the team expects to launch in 2027. Before that, the agents will also learn to do more than just think with you, Sesame hints, suggesting they’ll later be able to take action on your behalf — hence why they’re called “agents” in the first place, instead of just chatbots. That is potentially even more interesting, as working with agentic tools or apps today requires being able to prompt for what you need and have a specific idea of what you want to happen, and sometimes, even how it should happen. A conversational agent that you could talk to naturally could help you take the next steps, without you having to perfect the command you’re giving it. TheiOS appis out today in 39 countries, and the full experience is free for the time being. However, there still may be a short waitlist at sign-up. AnAndroid previewis coming in the future, the company says.
View
How long is Anthropic’s lease with SpaceX? Opinions vary.
Earlier this month, xAI signeda major compute deal with Anthropic, pledging billions of dollars a month for exclusive use of the company’s Colossus cluster. It was a coup for both companies, giving xAI some much-needed revenue and helping Anthropic catch up in the never-ending race for compute. Butthis morning on X, Elon Musk downplayed exactly how much SpaceX had committed to the deal. “SpaceX has not committed to leasing Colossus for years, although it’s possible that may be what happens,” he said, replying to a user. “This is a 180 day lease with 90 day notice mutual cancellation thereafter. The short term was our request, not Anthropic’s. We won’t leave them hanging and will provide a reasonable off-ramp, but if compute gets super tight I said we might need it back at some point.” Musk’s statement directly contradicts SpaceX’srecent S-1 filing, which confirms the standard 90-day cancellation but presents the deal as a three-year agreement. Page F-62 of the filing reads: On May 3, 2026, the Company entered into a cloud services agreement with Anthropic PBC, an AI research and development public benefit corporation, with respect to access to compute capacity. Pursuant to this agreement, the customer has agreed to pay a monthly fee through May 2029, with capacity ramping in May 2026 at a reduced fee. The agreement may be terminated by either party upon 90 days’ notice. The customer will retain ownership and intellectual property rights in its content, AI models, and related data. The key point here is that Anthropic “has agreed to pay a monthly fee through May 2029” — a pretty straightforward description of a three-year lease. The same language is repeated on F-96 and in slightly varied form (“the customer has agreed to pay us $1.25 billion per month through May 2029”) on pages 13 and 146, so it’s not as if there was a typo. xAI did not respond to a request for clarification. Maybe we can quibble about whether Anthropic agreeing to pay for a service means the same thing as SpaceX agreeing to provide that service, but that’s not usually what “lease” means. And why have a one-way lock-in if either party can terminate the deal with three months’ notice anyway? I don’t have the deal in front of me, so I don’t know what it says — and neitherSpaceXnorAnthropicis saying anything about the duration of the deal in their announcements. Still, there should be a pretty straightforward fact of the matter here, and it’s not the sort of thing you want to make false statements about during a company’s quiet period. As always, we should note that the SEC probably will not do anything — and even if they did, Elon probably wouldn’t care. But this sort of does seem likea material misrepresentation made while marketing a security, which is bad karma at the very least. Sean O’Kane contributed reporting to this article.
View
Anthropic releases Opus 4.8 with new ‘dynamic workflow’ tool
On Thursday, Anthropicreleased Opus 4.8, the newest version of its most advanced publicly available model. The model is available everywhere, with standard pricing at the same level as the previous Opus release. The new model comes just 41 days after Opus 4.7 was released, a much faster upgrade cycle than normal for Anthropic. (The most recent Sonnet and Haiku models are three and seven months old, respectively.) The fast turnaround may have something to do with the chilly reception to Opus 4.7, which some usersfounddisappointing. That interval has also seen significant new releases forOpenAI’s Codexand Google’sGemini Flash model, increasing the pressure on Anthropic to keep pace. Opus 4.8 comes with the expected best-in-class benchmark results, but there’s also particular attention to how the model manages bad or uncertain data. In the launch post, Anthropic’s early testers found that the new model is “more likely to flag uncertainties about its work and less likely to make unsupported claims.” Echoing this point, a testimonial from Bridgewater associates said the biggest difference in the upgrade was “Opus 4.8’s tendency to proactively flag issues with the inputs and outputs of an analysis, something other models routinely missed and left to the users to catch.” Together with the new model, Anthropic launched a feature calledDynamic Workflows, which will be available in research preview. The system is designed to help larger models like Opus manage complex tasks across hundreds of parallel subagents. “Claude Code alongside Opus 4.8 can now carry out codebase-scale migrations across hundreds of thousands of lines of code from kickoff to merge, with the existing test suite as its bar,” the post explains. Anthropic is still holding back its most advanced Mythos model aftera tentative preview last monthraised cybersecurity concerns. However, the company hinted in today’s Opus release that the Mythos preview period might soon end, once necessary safeguards are complete. “We’re making swift progress on developing these safeguards and expect to be able to bring Mythos-class models to all our customers in the coming weeks,” the company wrote.
View
In just 3 weeks, StrictlyVC is coming to Los Angeles
Join us forStrictlyVC Los Angeles 2026, an intimate evening bringing together leading investors and entrepreneurs for high-signal conversations from the front lines of venture capital and frontier technology. Taking place Thursday, June 18, at The Aerospace Corporation Campus in El Segundo, this edition continues StrictlyVC’s focus on direct access to the ideas and leaders shaping where technology and capital are headed next.Secure your spot here. For executives, investors, and founders navigating an increasingly complex market, this is an opportunity to step inside conversations that rarely happen in public and hear directly from the people driving change across defense, AI, and advanced industry. We will begin the evening withEthan Thornton, founder ofMach Industries. In his session called “Built for a New Era of Defense Technology,” Thornton will discuss what it means to build a hardtech company at speed and why defense innovation is undergoing a structural shift as autonomy, manufacturing, and national security become increasingly interconnected. His perspective reflects a new generation of founders choosing to operate in industries once considered slow-moving, now rapidly reshaped by technological acceleration. Next the conversation turns to backing the next frontier of physical AI, featuringDelian AsparouhovofFounders FundalongsideSaif KhawajaofShinkei Systems. Together, they will explore how advances in AI, robotics, and automation are beginning to reshape not just software systems but the physical world itself and what it takes to move breakthrough technologies from concept to real-world deployment at scale. Additional speakers and conversations will be announced in the weeks ahead as the StrictlyVC Los Angeles agenda continues to take shape.Stay updated on new speaker announcements and event developments. As the evening unfolds, the room becomes the real value of the event. Conversations continue beyond the stage in a setting defined by access, focus, and proximity to the people actively shaping the next generation of companies. It is an environment where introductions turn into insight, and insight often turns into opportunity.Secure your spot here.
View
Devin Maker Cognition Raises $1 Bn at $26 Bn Valuation as Adoption Grows
The company said its annualised revenue run rate has reached $492 million.
View
Has the hunt for AI compute uncovered the next Cerebras?
The raging demand for computers to run AI models has only accelerated, but there are two major obstacles that anyone in the business needs to overcome: getting the right chips, and getting them into data centers where they can start generating revenue. General Compute, a new inference neocloud — a company that rents out AI processing power, specializing in the phase when models are running and responding to users rather than being trained — has answers to those questions that illuminate where the AI ecosystem is headed. Those answers helped it raise a $15 million seed round at a $60 million post-money valuation, led by FUSE VC with participation from Carya Venture Partners and Village Global Ventures. First, what is the right chip? The demand for GPUs has gone through the roof, but it’s becoming conventional wisdom that they aren’t the best-suited chips for running AI models once they have been trained. The phase of AI where a model is actively generating responses has different computational requirements than training, and a new class of chips is being designed specifically for it. Nvidia’s $20 billion Groq transaction in December and Cerebras’ $57 billion IPO last week point the way. With capacity strained at both those companies, the co-founders of General Compute, CEO Finn Puklowski and CTO Jason Goodison, found another option. They’re turning to specialized chips built by SambaNova, an Intel-backed chipmaker focused on inference that has fallen a bit out of the Silicon Valley conversation. That may change when SambaNova releases its new chips this year. The architecture is more flexible and uses more memory to store context during inference calculations, and SambaNova claims that it outperforms not just GPUs but also other specialized chips built by the likes of Groq or Cerebras. Puklowski says the new chips will generate 600 to 700 tokens per second, versus about 250 tokens per second for GPUs. General Compute has $300 million of the company’s SN50 chips on order and says it will be the first neocloud deploying them. These chips also help solve the second big problem—where to put them—for General Compute: They are air-cooled, not water-cooled, and consume less power, so they can be installed in existing data center facilities without new infrastructure investments. Puklowski is pursuing colocation deals — arrangements where General Compute installs its hardware in someone else’s facility — not just with data center providers, but also with crypto miners looking to repurpose their infrastructure as the cost of producing a bitcoin has often exceeded its price. General Compute launched its cloud offering last week, claiming it is already the fastest at running MiniMax 2.7, a powerful open-source LLM. Joe Hasselmann is a venture investor who got in on the ground floor of the inference boom when he invested in Groq in 2021. This year, he launched a new fund, Evercrest Capital Partners, focused on the AI space, and made General Compute his first investment. Hassleman sees in SambaNova’s partnership with General Compute parallels to Coreweave’s relationship with Nvidia — and to the pairing of Groq’s chip-making with its former cloud offering. “They do need a healthy mix of customers that are going to put their chips in environments that are going to have high growth to them,” Hassleman said. “As much as General Compute is making a bet on SambaNova, SambaNova is making a bet on General Compute.” The question is what kind of computer architecture will capture the most value in the AI future. Inference clouds are implicit bets on a world of multiple models and agents, one where no single provider dominates and speed and cost of inference become the key competitive variables. Consider the$113 million Series Braised for OpenRouter this week, reflecting the company’s ability to offer customers access to multiple models in order to optimize their token spend. Speed matters in that calculation, for price, and for capability. Puklowski wants to turn hour-long workloads for coding agents into five- or ten-minute tasks, and make audio agents for customer service, which require faster inference to converse effectively, more economical.“If you use ChatGPT and it gives you 50 tokens per second, that’s still a heck of a lot faster than we can read,” Puklowski told TechCrunch, “Now that things have moved to agent-to-agent, where agents are out there reading on our behalf or pinging databases, they need to go faster.”
View
Confluent Launches Dedicated GCC Strategy in India as AI Adoption Accelerates
the GCC ecosystem is undergoing a structural shift, with centres functioning as captive support units
View
Wipro Expands ServiceNow Partnership to Scale Agentic AI Workflows Across Enterprises
Wipro Intelligence suite will be integrated with ServiceNow’s AI platform to automate workflows across functions such as IT, HR, procurement and cybersecurity.
View
Google Partners with Polaris School of Technology to Launch AI and Cloud-Focused Degree Programme
Google collaborates with Bengaluru's Polaris School of Technology to integrate certified pathways, AI tools, and multi-million-rupee credits into a new degree.
View
TCS Partners Mistral to Build Custom Enterprise AI Models at Scale
TCS has partnered French AI startup Mistral to help enterprises build custom AI models using proprietary data.
View
Why India Inc Struggles to Retain Gen Z in the AI Economy
India’s employers are facing a new workforce reality as Gen Z workers, armed with AI-era skills, are moving faster than traditional career ladders can keep up with.
View
Is Claude Mythos Overwhelming Security Teams?
Organisations can find more vulnerabilities than ever before, but the sheer volume forces them to fix their fundamentals first.
View