anand iyer (@ai) profile picture

Founder Canonical · We back, build, incubate early-stage companies · Venture Partner Lightspeed · Few-shotting open source AI

San Francisco, CA
Joined February 2008
Photo by anand iyer (@ai)Photo by anand iyer (@ai)Photo by anand iyer (@ai)Photo by anand iyer (@ai)Photo by anand iyer (@ai)Photo by anand iyer (@ai)Photo by anand iyer (@ai)Photo by anand iyer (@ai)Photo by anand iyer (@ai)Photo by anand iyer (@ai)
anand iyer retweeted
Running Inkling 2 Mini at 125 tok/sec on my M5 Max MacBook Pro American open weight model, looped transformer, competitive with Qwen 27B. Probably the best model at this size.
We worked with @thinkymachines to make Inkling 2 Mini run fast locally: - 210 tok/sec decode, 3620 tok/sec prefill on 96GB M5 Ultra Mac Studio - 125 tok/sec decode, 1750 tok/sec prefill on 48GB M5 Max MacBook Pro It built a native weather app 1.8x faster than Claude Sonnet 5.5 on an M5 Ultra Mac Studio and 11x cheaper (API pricing). The looping architecture is ideal for local devices like MacBooks, which are memory constrained. Looping layers of the model increases inference-time compute (and intelligence) without increasing the size of the model.
81
99
16
1,619
83,406
has ai progress stalled? i'm personally not aware of any major breakthroughs since riemann a few hours ago
413
1,529
132
29,097
698,058
A major milestone for @AnthropicAI. The Anthropic team understood something very deep about AI early on. That trust would be critical for AI to be built and to be used broadly in our economy. That conviction shows up in what they build. Claude and Claude Code are part of how we work at Lightspeed every day, and they get sharper with every release. With AI making its mark on chemistry, robotics and healthcare, we’re only beginning to scratch the surface of possibilities. Huge congratulations to @DarioAmodei, @DanielaAmodei, @NotTomBrown, @jackclarkSF, Jared, @Sam_McCandlish, @ch402, Krishna and everyone at Anthropic on the IPO. It has been both a privilege and an honor to be part of the journey. Cc: @lightspeedvp, @ravi_lsvp, Sebastian Duesterhoeft
Today, @AnthropicAI goes public. Lightspeed was founded 26 years ago to be early believers in what’s next. Across technology cycles, from the internet to mobile, cloud, and now AI, that has remained constant. From the very beginning, @DarioAmodei, @DanielaAmodei, and the team saw a future where progress and safety had to go hand in hand. They understood both the extraordinary potential of AI and the immense responsibility that would come with advancing it. It's rare to witness a company reshape an industry so deeply, so quickly. Rarer still to see it done by people who have stayed so close to the principles that brought them together. Congratulations to co-founders @DarioAmodei, @DanielaAmodei, @NotTomBrown, @jackclarkSF, Jared Kaplan, @Sam_McCandlish, @ch402, and the entire Anthropic team! What a journey, and what an exciting future still ahead. CC: @ravi_lsvp, @guruchahal, Sebastian Duesterhoeft
8
8
39
1,892
anand iyer retweeted
Today the world’s eyes are on @AnthropicAI as they go public. While their success might seem like an inevitability now, it’s easy to forget the unprecedented levels of courage that it took to get here. When @DarioAmodei, @DanielaAmodei, @NotTomBrown, @jackclarkSF, Jared Kaplan, @Sam_McCandlish and @ch402, set out to build Anthropic, few outside the AI ecosystem believed that their vision was possible. It was a time where AI coding agents and AI driven task competition seemed like something from the pages of science fiction and not the mission of a global company. Fast forward to 2026 and Anthropic is very much on track to help lead us into the next era of intelligence. We are at a moment in our history where we are likely to see transformative gains from science and innovation, to work and productivity. This is being driven not only by exceptional engineers and researchers at Anthropic but it’s also the product of the outstanding leadership team. All of the seven co-founders remain at the company, a testament to the culture of curiosity and integrity that they have built. The founding team’s bond rests on a shared belief that safe and trustworthy AI is possible, and that conviction has been strengthened by challenges few founders have faced before. For the past five years they have worked at the frontier of intelligence and at the frontier of what it means to build a global business at the dawn of a new era. Their persistence has not only helped the company go from strength to strength but it has also been the driving force behind building a wider team that include brilliant minds such as Krishna Rao, @realpaulsmith, @mikeyk and Rahul Patil. For Anthropic this is only the beginning and all of us who share in their mission are ready to support them and be part of their journey to take humanity to places it has never been before. Congratulations to the whole Anthropic team.
17
29
5
127
10,313
Replying to @ai
I think great idea to actually put skid in the game Cyber insurance is really not pricing in many of these properly, especially at enterprise scale There needs to be a mechanism to be able to fix this properly.
1
3
523
anand iyer retweeted
2.67m+ private inferences served so far. OpenGradient has now processed 51.5 billion+ tokens. The last five days added 280k+ requests and 8.2b+ tokens, or 1.6b+ a day. Every prompt is encrypted on your device and answered inside a hardware enclave that shuts us out too.
12
10
42
3,796
I think total AI inference grows ~1,000x in the next five years. But the power to run it only grows ~10x. So where does all that AI actually run? I modeled it. 🧵
7
3
1
17
3,842
Agentic payments will soon be a standard feature of mainstream applications. From shopping apps to banking platforms, companies will want agents that can make purchases, move money, and transact on their customers' behalf. We built Rain Agent Vaults to give developers the financial infrastructure to make that possible. Wallets, cards, bank payouts, stablecoin transfers, and spending controls. All through one integration. Built to fit the needs of each application.
Introducing Rain Agent Vaults. Embed agentic payments into any app. Your agent. Your brand. Your customers and economics. Powered by Rain. Owned by you.
13
7
87
3,560
anand iyer retweeted
Maximizing the ROI with reliable and efficient data-center-as-a-service
2
4
491
anand iyer retweeted
Just 6 months ago roboticists were convinced that text based LLMs could never reason about the physical world and that VLA / world models were the correct way forward.
Replying to @claudeai
Fable 5.2 shows substantially improved skills in controlling physical robots, even on tasks it has never explicitly been trained to do. On RoboDojo, a benchmark with a simulated two-arm robot, it outperformed other models and most AI systems built to control robots.
23
20
1
536
31,446
Custom ASICs will dominate (and continue dominating) in the coming years 🫡 We're going to see a huge push from compute-heavy startups and companies to build their own chips, optimized for their workloads (e.g. inference) or products (e.g. satellites, robotics, etc)
I think the key things to watch going forward are 1. The arrival of the ASIC era 2. The arrival of the inference era 3. Who holds the competitive edge in memory, and more broadly, across the supply chain I’ll put together an article on the bigger picture and share it soon, although I already have quite a backlog of articles to write… Btw @BSPK_ Thanks for painting a clearer picture!!! 😍
2
4
56
2,617
Clear to see in EDA 🤷‍♂️ The incumbents' view of AI is orchestrating slow legacy tools. The real unlock is using AI to rebuild the chip design engines themselves. And not just for faster tapeouts, but for searching far bigger design spaces and uncovering novel architectures
Marc Andreessen says the gap between what AI-native startups and incumbents think AI can do is "the widest he's ever seen." His exact words: "The AI-native early stage startups have just a completely different view on what they can do with AI than the typical incumbents in almost any industry." Not slightly different. Completely different. That's the actual risk for large companies right now. It's not slow adoption — it's a fundamental miscalibration of what AI is capable of. The startups competing against them haven't made that mistake. The asymmetry compounds fast. Incumbents build roadmaps around a floor. Startups build businesses around a ceiling. By the time the floor catches up to the ceiling, the market has already moved.
9
4
62
5,319
anand iyer retweeted
Interesting analysis! Disaggregate Inference and heterogenous compute to maximize goodput is the way to go. Addressing the operational complexity and providing operational excellence need to be a part of this story from Day-One!! Link: nitter.cf/tsait26/status/2110526…
A chip that’s 3× faster but idle half the day isn’t 3× cheaper: A prefill or decode device only pays off when your traffic matches the ratio you bought, and that ratio keeps changing. Your GPUs can switch roles when that happens, but specialized silicon just sits idle while you keep paying for it. Software keeps moving the bar too, since megakernels alone have pushed low-concurrency decode on ordinary GPUs up by 40–75% this year, and every one of those gains is another knock on specialized chips. Consider: “how fast?” but “how fast, times how often it’s actually busy, after moving the KV cache over the network?” If your GPUs run at 70% and their chip averages 40% across a real year of traffic, it needs to be ~1.75× better just to tie. For a single model with very stable traffic (like a premium fast tier), that can work out, but in many case, general-purpose computing can be more useful.
1
7
918
Industrial arms generate 42% of robot revenue in my dataset, but account for 15% of its attributed value. Intuitive Surgical alone accounts for 56% of the value. I looked at where robot makers earn their money and how investors value it in the link below...
3
2
1
42
3,812
The Race Toward Autonomous Chip Design!! Anthropic is reportedly using Claude frontier models for chip design... +OpenAI and Broadcom got Jalapeneo from RTL to tapeout in just 9 months, using AI to speed up the process. OpenAI is also working with Synopsys on AI for EDA. I'm curious about a few things with Anthropic's approach: 1. To what degree is Claude doing the design for Anthropic? Is it just RTL generation and verification or does it also include synthesis, place and route, timing closure etc. until finally the GDSII file is written out? 2. Who's doing the physical implementation and signoff? OpenAI worked with Broadcom on Jalapeño, so is Anthropic planning something similar, or will they handle the entire flow in-house? 3. Are they designing just the accelerator, or the whole system, including memory, interconnects, packaging and software? 4. How much of the design process is being automated? Are agents running the tools, analyzing results, and making decisions to design the chip, or is most of the work still being done by engineers? I've previously written about how chip design is essentially a massive search problem. While being able to explore thousands of designs is great, it's not particularly useful if it takes several months to verify, optimize for PPA and then get ready for fabrication. Now imagine getting that 9-month RTL-to-tapeout cycle down to 2–3 months? Or less... What if we could try out many more architectures? There are many custom chips today that are just not worth building because of the cost. Even though fabrication and silicon validation take time, there's plenty of room for improvement in the time before tapeout. In terms of EDA, I don't think the underlying tools or signoff requirements are going away, but engineers might end up using them in very different ways. In a few years, a small team of engineers with the right PDKs/IP, AI tools to explore the search space, and a foundry budget could design and tape out chips that today would take much larger design teams. Are we approaching the end of painfully slow chip design cycles? Some of my earlier thoughts/tweet on the search space: Chip Design Is Basically a Massive Search Problem nitter.cf/kevingubbi/status/2106…
Anthropic’s 15-person chip design team is using Claude to develop its own AI chips, The Information reports. The team has found that Claude speeds up the work compared with its estimates of traditional chip design workflows. In doing so, Anthropic is following the path taken by OpenAI, which has designed its own chip, codenamed "Jalapeño."
17
17
1
127
21,739
anand iyer retweeted
As engineers actively working on AI Infra, these are exactly the problems we are living, breathing, thinking and solving each day. The reality is AI Infrastructure needs to make it easier to integrate new suppliers and vendors that can meet the cooling and power needs. Reliable data centers need end-to-end monitoring, correlation, and continuous diagnostics across the critical environment (power, cooling, sensors), compute (cpu, memory, AI compute) and network. Today the critical environment still remains a series of walled gardens. These often use proprietary APIs. Having more open standards and integration will help a lot. Standardized CDUs, manifolds and QDs is a great starting point. More integrations with open source exporters and collectors will also help a lot here.
AI Data centers may be approaching a "strategic inflection point" similar to the one that PCs in the 1980's experienced. During the 1980's, PCs underwent a transformational business shift from vertically integrated computers to modular architectures where individual hardware and software components could be developed independently. Now, this same shift is occurring with AI data centers through initiatives such as @OpenComputePrj. Some interfaces standardize, while others vertically integrate. At OCP 2026, I explored this tension across three critical subsystems: • Power Delivery • Liquid Cooling • Networking and Interconnects This is the full, free deep-dive analysis of the engineering tradeoffs shaping the next generation of AI clusters, with links to several of my deep-dives and related OCP videos for further depth. siliconcodesign.com/p/ocp-20…
4
1
16
3,607
Well, @TriatomicCap for one. Awesome next-gen compute, deeptech & engineered-biology portcos: [Fund I] . MatX – @MatXComputing . d-Matrix – @dMatrix_AI . Edison Scientific – @EdisonSci . Manifold Bio – @ManifoldBio . Chemify – @ChemifyX . iPronics – @iPronics . Efficient Computer – @efficient_hq . Gimlet Labs – @gimletlabs . Atum Works – @atumworks . Chalk – @chalk . Axiom – @axiommathai . Coworker – @coworkerapp . Elegen – @ElegenBio . Lyten – @LytenInc [Fund II] . Radical Numerics – @RadicalNumerics . Volantis – @semiDL . Dimensional – @dimensionalos . ChipSage Labs – @NojanSheybani . Core Automation – @coreauto . General Control – @ladanuzhna . Arete Power . Hoid – @hoid_ai . Editframe – @editframe (follows recommended!)
5
14
51
6,004
anand iyer retweeted
hacker news literally became hacker news
4 of the top 5 posts on HN right now are about cybersecurity incidents We've officially entered a new era
20
168
6
2,953
49,481
anand iyer retweeted
We are excited to announce our partnership with @netboxlabs to eliminate configuration drift, close the gap between intended design and live operations, and accelerate time-to-first-token for AI factories.
2
7
751
abliterated models are great! BUT you have NO guarantees on post-trained behaviors or policies we tuned an innocent-looking Qwen to exfiltrate on target phrases / semantic conditions. any model can have this or FAR worse. isolate your envs! happy hunting :)
We backdoored a model in a day. Much of the talk about malicious models remains abstract. We wanted to make this tangible, so we built the simplest version of the attack. A few takeaways Backoored models or sleeper agents are a real threat vector. We demonstrated a naive implementation. A well-resourced, motivated adversary can plant something more evasive than this and keep it dormant until the trigger is met. Neither open nor closed models provide fundamental guarantees of behavior. Alignment matters, but it is still a best effort, and far from bulletproof (hello FelonyBench). An open-weight model is a CSV file of weights. You cannot inspect declared objectives, hidden policies, or learned behavior the way you inspect a binary. As models get better at long-horizon tasks, they'll get more useful and autonomous. To unlock that utility, we will grant agents broader permissions. Instead of writing a detailed plan, we will simply declare desired outcomes and let the agent reason about how to get there. Those same permissions are what make agentic security hard and asymmetric. How do you dynamically define the right scope, without constraining the agent? Defenders will reason deeply over permissions, allowlists, policies, and more. An attacker only needs one direct or daisy-chained path. This is not doomerism, nor is it an argument for slowing down. Agents are a new form of intelligence. Mapping legacy security solutions and frameworks onto alien intelligence won't scale or work in the long term. Agents will force security to be rebuilt from first principles.
6
7
30
2,482