Keeping up with AI shouldn’t feel like a full-time job. Forward Future delivers the most interesting stories—curated, concise, and straight to your inbox every morning at 7 AM ET. Stay informed without the overwhelm.👇 https://t.co/Ieba4mt5cz

If agents create lock-in, do we need an open-source or open protocol similar to MCP? During Google I/O, we asked Sundar Pichai: “Agent memory is what makes agents dramatically more powerful. When they learn shorthand with you, quality and efficiency jump.” “But memory also

AI hasn’t changed the internet as much as X thinks it has. The CEO and co-founder of @Amplitude_HQ on the real impact of agents on traffic and websites: “You see some impact to Google traffic and some impact to websites, but for the vast majority of businesses, usage patterns

Big show tomorrow at 10am PT on Forward Future Live. Featuring: • Doug O’Laughlin (President, @SemiAnalysis_) • @koraykv (CTO, @GoogleDeepMind) • @abhishekf96 (VP Product, @RobinhoodApp) Plus a late addition: @OfficialLoganK breaking down the Gemini 3 Flash release. See you

The real moat when AI kills most SaaS UIs. @spenserskates CEO & cofounder @Amplitude_HQ on where software is headed next: “ChatGPT and similar interfaces are going to completely disintermediate the first generation of SaaS tooling.” He echoes a framework Satya Nadella has

Will voice be the dominant AI interface? Maybe not. Spenser Skates (@spenserskates) on why the jury is still out on voice-first computing: “It still hasn’t replaced the keyboard, mouse, or touch screens. The jury is still out.” The core issue is feedback and latency: “With

Benchmarks as a forecasting tool. Tejal Patwardhan (@tejalpatwardhan) from @OpenAI on why evals matter beyond ranking models: “They’re useful as a forecasting tool, they tell you how fast AI progress is actually happening.” “Sometimes you build a benchmark thinking you have

AI keeps unlocking problems we couldn’t solve before. @trustvanta’s CPO Jeremy Epling on what’s actually changing inside the product as models improve: “There are a lot of areas where we were trying to hit a quality bar and just couldn’t. Then a new model drops, and suddenly

"The biggest risk for Gemini is running out of innovation." Great conversation with @OfficialLoganK and Google’s Chief AI Architect and DeepMind CTO @koraykv on the release of Gemini 3. On Google’s "Underdog" Journey: "I’ve been very honest with the team... we were nowhere near

"We now need to be fighting agents with agents." @TrustVanta CPO Jeremy Epling says the security landscape has fundamentally shifted. "You have attackers leveraging agents... raising the bar on what internal security means." It is no longer human vs. human. It is AI vs. AI.

The billion-dollar question: In a world where everyone has AI, what actually helps you stand out? Fabian Stelzer (@fabianstelzer) says the answer is still human creativity: “At the end of the day, it comes down to human creativity. The models aren’t super creative, and even

“The skepticism was that the button didn’t exist.” @christinacaci CEO and co-founder of Vanta, on convincing early customers that software could replace manual compliance work. Customers told her: “If there were a button I could press to get the same outcome — easier, faster,

The biggest AI gains don’t always come from better prompts. @jeremy_epling CPO @TrustVanta on what’s actually driving performance improvements right now: “Models are getting more diverse and more specialized. Finding the right model for the task matters more than ever.” “We’ve

If you’re not leveraging AI, you’re already behind. @spenserskates CEO & co-founder of @Amplitude_HQ, on what AI is already doing inside real teams: “What it can do is speed up existing human work, a lot.” “If you’re an engineer or in support and you’re not leveraging AI,

From “worse than a human” to “preferred most of the time.” @tejalpatwardhan researcher @OpenAI on what changed with GPT-5.2: “Before, the model would lose to a human expert most of the time. You’d only prefer it about 38.8% of the time.” With 5.2, that flips: “Now you prefer

Star Wars is about to look very different. @fabianstelzer lays out how major IP could evolve in the age of generative media: "The future Star Wars might not be a long movie you watch. Instead, it might be a set of guardrails defined by the IP holder." Studios define the

"SOTA on basically every benchmark we care about.” @tejalpatwardhan from OpenAI on GPT-5.2: “5.2 is a really big deal. Calling it a 0.1 increment understates how excited I’ve been about this model.” “It’s our best model on enterprise work by far, and it’s SOTA on basically

We’ve got a good show for you this week. Here’s the lineup: - Doug O’Laughlin (President, @SemiAnalysis_) - @koraykv (CTO, @GoogleDeepMind) - @abhishekf96 (VP Product, @RobinhoodApp) Join @MatthewBerman + @NickWentz_ Friday at 10am PT.

Traditional benchmarks don't tell the whole story. Tejal Patwardhan (@tejalpatwardhan), researcher at OpenAI, explains: “A lot of the capabilities that matter in an economic setting weren’t really being measured by our benchmarks.” Benchmarks like GDPval are an important next

What the Disney–OpenAI partnership means for film. @fabianstelzer, Co-Founder of @heyglif, says “Someone will come up with a great idea in Sora for a Disney show, and then they’ll make that show. That’s going to happen.” But that’s only the beginning. “Long term it’s going to

What if you could 𝙩𝙚𝙖𝙘𝙝 an AI instead of just prompting it? 🤔 Our #RTXAIGarage blog explains how 𝗳𝗶𝗻𝗲-𝘁𝘂𝗻𝗶𝗻𝗴 𝗮𝗻 𝗟𝗟𝗠 works and why NVIDIA RTX GPUs & DGX Spark make it easy to do locally using @UnslothAI. 🔗: https://t.co/I086DnEv3r