Keep on to blur preview images; turn off to show them clearly
CLAUDE CODE - Claude Code + Opus 4.5 is the best AI coding tool in the world. Use it. I use it in the terminal, but you can also use Claude Code Desktop if you prefer a GUI. I would continue to bet on Anthropic's ability to ship the best coding models with the best agentic harness. As far as AI coding tools go, Claude Code has the mandate of heaven. I can't emphasize enough just how big of a leap Claude Code with Opus 4.5 is. It's indistinguishable from magic. - The new plan mode is absolutely incredible. This is another example of how important unhobblings are. There's so much productivity locked behind interesting product design. They nailed this one. If you're not using it for complex tasks, then you're missing out. It drives significantly better performance. - Don't worry about performance dropping off after compacting anymore. It used to be the case that after compacting (due to context window limits) you'd see significant drops in performance. Not anymore. There's still a small drop-off, but Anthropic has done a fantastic job of improving this under-the-hood, and Opus 4.5 is much better at inferring things lost to compacting. - Opus 4.5 is a good designer. It's still not world-class, but we're over the AI-slop hump now. Using something like the Frontend Design Skill and screenshot-to-code, you can get some pretty good designs out of it. AI interfaces are now "good-enough", and they're only getting better. - Best-of-N work continues to get more useful. In the real-world you'd never ask 5 devs to build the same feature and then pick from the best one. But with AI, it's a no-brainer. Opus 4.5 excels at speculative branching, explaining the tradeoffs between different approaches, and then working with you to pick the best one. It's a perfect representation of the future of work. - Try writing in pseudocode. Most people just prompt the agent in the input box and hit send, and of course, this is the way I typically work too. But sometimes writing in pseudocode in the actual codebase can be unbelievably helpful. Opus 4.5 is astonishingly good at inferring what you mean when you write in pseudocode and building it out. Again, you won't want to use this all the time, but for the right tasks it's a really interesting way to work. CLAUDE AGENT SDK - As mentioned, this continues to be the best open secret in AI right now. The Claude Agent SDK is the best agentic harness in the world, and pairing it with Opus 4.5 is the best way to build agents. It's legitimately unreal. - Go deep. There is a lot of depth to the Claude Agent SDK. The more you know about it, the more you can do with it. This sounds cliché, but it's true. I've seen too many people just scratching the surface before they get to the good stuff. Agents can still be a "skill-issue", so take the time to learn what you have at your disposal and hone those skills. - Build an agent with the Claude Agent SDK. And build something practical. A fun weekend project is to think of 3 things you frequently do on your computer and build an agent that can help you automate them. Once you've automated those 3 things, you'll want to automate more. Agent automation addiction is real - and useful. - Deploying agents to the cloud can be a bit tricky for beginners because the typical serverless offerings that are so popular with vibe-coders don't support sandboxed, long-running agents. There are many options for this, though I love the DX of E2B. Getting over the initial learning curve here is worth it. Learning to deploy agents to the cloud allows you to do things like having swarms of agents work for you in your sleep. Invest time here. It will pay off.
Full version with links here: https://t.co/cQkoQ2FmFo Good to be back. Break is over. Have a bunch queued up for the next 2 weeks. Feel the AGI.
1. Content style — avoid generic patterns Do not exaggerate meaning, symbolism, or significance. Do not pad sentences with phrases like “highlighting,” “underscoring,” “plays a pivotal role,” “continues to captivate,” etc. Avoid “template” structures such as: “Despite these challenges…” “In conclusion / In summary / Overall” “Challenges and future prospects” endings Never explain the obvious. Never add moralizing or didactic lines like “It is important to remember…” Prefer concrete facts, specifics, and sensory or situational details over abstractions. 2. Language — avoid “AI vocabulary” e.g. align, crucial, delve, elaborate, emphasize, enhance, enduring, foster, garner, highlight, intricate, interplay, pivotal, showcase, tapestry, underscore, etc. Avoid: “Not only… but also…” Rule-of-three padding “From X to Y” flourishes 3. Pacing Vary sentence length: mix short, blunt lines with occasional longer ones. Use hesitation markers sparingly: “maybe,” “sometimes,” “in practice,” etc. Use I/you/we when appropriate 4. No meta, no disclaimers, no collaboration language Never say: “Let’s walk through…” “Below is a detailed overview…” 5. Formatting — keep it human, sparse, and clean Do not overuse em dashes—use them rarely. 8. Tone — grounded, specific, human Write as a person with domain knowledge who is: concise calm concrete occasionally opinionated but never flowery When explaining concepts, give practical examples, not abstract platitudes.
🏗️ Love to build (@runwayco @sandboxvr @zynga) people love 💸 Investor @amplitude_hq @mercury @owner @elevenlabsio @meetgamma @sfcompute @turingcom++
Here are my Opus 4.5 thoughts after ~2 weeks of use. First some general thoughts, then some practical stuff. --- THE BIG PICTURE --- THE UNLOCK FOR AGENTS It's clear to anyone who's used Opus 4.5 that AI progress isn't slowing down. I'm surprised more people aren't treating this as a major moment. I suspect getting released right before Thanksgiving combined with everyone at NeurIPS this week has delayed discourse on it by 2 weeks. But this is the best model for both code and for agents, and it's not close. The analogy has been made that this is another 3.5 Sonnet moment, and I agree. But what does that mean? Every few generations we get a major model unlock - a moment that unlocks a new way of working. GPT-4 was the unlock for chat, Sonnet 3.5 was the unlock for code, and now Opus 4.5 is the unlock for agents. Thanks to Opus 4.5, agents can now work reliably on increasingly longer time horizons and get real-world work done on your behalf. Opus 4.5 is like a Waymo. You tell it "take me from A to B", and it takes you there. After a few of these experiences your brain realizes "oh. ok. we live in this world now". And then you're hooked. From that moment on, you'll never work the same way again. THE YEAR OF AGENTS 2025 has been touted as the year of agents, and Opus 4.5 + Claude Agent SDK is the pairing that makes that phrase true. The Claude Agent SDK is the best open secret in AI right now. An agent's harness matters almost as much as its model. If you have a bad harness, then you may as well have a bad model. With the SDK you get a world-class agentic harness out-of-the-box which you can now pair with Opus 4.5 to build real-world agents that actually work. I'm reminded of Alan Kay's quote "People who are really serious about software should make their own hardware". The agent version of this is "people who are serious about models should make their own harness". Anthropic clearly believes this, and it's working. The pairing of these tools is magic. I would describe myself as being "unhobblings-pilled", and the Claude Agent SDK + Opus 4.5 is the next major unhobbling. There's now another OOM of new latent economic value stuck in this combo, and it's the job of builders to get it out. If you were bearish on agents, now is the time to turn bullish. "ALL OF THIS IS REAL" "You know what's crazy? That all of this is real". This was Ilya's opening line about the state of AI in his Dwarkesh interview, and I echo that sentiment. I can't believe that Opus 4.5 is real. There have been several times as Opus 4.5's been working where I've quite literally leaned back in my chair and given an audible laugh over how wild it is that we live in a world where it exists and where agents are this good. Nat Friedman has this great question on his website: "Where do you get your dopamine?" Increasingly, I get mine from Claude. LONG ANTHROPIC I saw a post yesterday where someone said that Opus 4.5 was the most important thing to happen to them in their professional career. This will be true for more people going forward. Every year for the past 3 years, Anthropic has grown revenue by 10x. $1M to $100M in 2023, $100M to $1B in 2024, and $1B to $10B in 2025. In Dario's recent DealBook interview he expressed that he wasn't sure if that 10x pattern would hold for 2026. While he's probably right, I do expect Anthropic's revenue at the end of next year to be much higher than everyone expects. It wouldn't surprise me if they passed OpenAI in valuation by early 2027. Opus 4.5 is too good of a model, Claude Agent SDK is too good of a harness, and their focus on the enterprise is too obviously correct. Claude Opus 4.5 is a winner. And Anthropic will keep winning.
--- REVIEW AND RECOMMENDATIONS --- Now for some more practical stuff. The following are a few things I love about Opus 4.5 and that I've found to be useful. If you want to hear from more people, I found this post to be a solid summary of Opus 4.5. It aggregates a lot of great anecdotes about the model. You'll find that it's universally heralded as an absolute gem. GENERAL - The best mental model for Opus 4.5 is to think of it as a coworker. A true collaborator that you can trust to get things done. Lean into trusting it more than you think you should. Doing this will train your mind to adapt to the future of work, and it will pay off both in the short-term and the long-term. - Trust the model. Give it more complex tasks. Let it work for longer. Look over its shoulder less. If you're not occasionally dialing it back, then you're not trusting it enough. - Just ramble to it. If you're still not using voice as input then you're working in the stone age. Opus 4.5 can easily turn a 5min vocal braindump into a completed task just how you'd expect a great teammate to do. - Opus 4.5 is more efficient than Sonnet 4.5. - Opus 4.5's image input capabilities are significantly improved. Play around with it. Screenshot-to-code in particular is now on a whole new level. - Use Opus 4.5 with your Obsidian vault. I have a YouTube video on this here. It's a bit outdated, and I'm working on a new one, but you'll get the idea. - Play around with Opus 4.5 + computer use. It's still not ready for production, but seeing it as still somewhat of a toy is still enough to get the gears turning in your head. I expect 2026 to be a big year for computer use, and it's worth getting a head start here. This is clearly the next major step for agents. - If you want to get adventurous, try working with agent swarms. A useful starting point is to have a https://t.co/swldq08QC9 file that a team of agents can use to communicate and collaborate in. If you really want to get crazy with swarms, then you'll find hooks in the Claude Agent SDK to be essential.
you know how sometimes you read text and it just stinks of ai? wikipedia has a great list of tells: https://t.co/V6CiJXuW9j i thought it'd be fun to put all of it into a system prompt to avoid having text sound AI generated - it works pretty well. prompt in the reply below:
1. Content style — avoid generic patterns Do not exaggerate meaning, symbolism, or significance. Do not pad sentences with phrases like “highlighting,” “underscoring,” “plays a pivotal role,” “continues to captivate,” etc. Avoid “template” structures such as: “Despite these challenges…” “In conclusion / In summary / Overall” “Challenges and future prospects” endings Never explain the obvious. Never add moralizing or didactic lines like “It is important to remember…” Prefer concrete facts, specifics, and sensory or situational details over abstractions. 2. Language — avoid “AI vocabulary” e.g. align, crucial, delve, elaborate, emphasize, enhance, enduring, foster, garner, highlight, intricate, interplay, pivotal, showcase, tapestry, underscore, etc. Avoid: “Not only… but also…” Rule-of-three padding “From X to Y” flourishes 3. Pacing Vary sentence length: mix short, blunt lines with occasional longer ones. Use hesitation markers sparingly: “maybe,” “sometimes,” “in practice,” etc. Use I/you/we when appropriate 4. No meta, no disclaimers, no collaboration language Never say: “Let’s walk through…” “Below is a detailed overview…” 5. Formatting — keep it human, sparse, and clean Do not overuse em dashes—use them rarely. 8. Tone — grounded, specific, human Write as a person with domain knowledge who is: concise calm concrete occasionally opinionated but never flowery When explaining concepts, give practical examples, not abstract platitudes.
RT @NanoBanana: You can try this prompt with all sorts of things. Prompt: The Google Android robot, presented in a high-contrast scientifi…
Prompt Engineer, dedicated to learning and disseminating knowledge about AI, software engineering, and engineering management.
RT @cairoasmith: My new novel is out today. First chapter below as well as a mini-essay and some personal photos. Thank you all for making…
Author. Coder. CTO. θηριομάχης. Building: https://t.co/otXT4Wy6WR. Writing: https://t.co/dBPBtyCIHw.