Nishaanth Reddy @reddmachine

Thrilled that Ubiklaw: Local Voice-to-Dashboard AI got featured in the @AITinkerers Community Spotlights! See it here: post-training.aitinkerers.org/p/top-ai-demos…

Today is a very historical moment for AI video generation You can now generate AI video faster than you can watch it Before it'd take let's say 2-5 minutes to generate 15 seconds of video @fal made a post-trained Minimax H3 variant called Max which is 50x faster than the original but still maintains quality It generates 15 seconds of video in 9 seconds! That means you can now do new things like build a perpetual livestream with it that never ends!

@

57 years ago

0 0 0 0 0

Qwen3.8-27B on one DGX Spark hit 36.59 coding tok/s with SGLang + DFlash2, 2.89× plain decoding. The fastest tool-safe run was vLLM + DFlash2 at 47.41 median tok/s with 18/18 native calls. 12 deployments. Full results and failure notes: morethanamachine.com/posts/qwen3-8-…

@johnennis Exactly this! There has to be an obvious reason these massive AI companies are failing at basic backward compatibility and QA testing.

I really didn't expect Codex to close the gap so quickly. The difference between what Codex was 3 months ago to right now is staggering! I’ve moved most of my day-to-day personal coding from Claude Code to Codex. Claude Code is still excellent, but it seems like for the first time in a long time, it has to play catch up. Kudos @OpenAI and @thsottiaux! It's more that they've stacked a few different wins at the right time. Anecdotally Codex with GPT 5.6 Sol as compared to Claude Code with Fable feels better in the following ways: 1. Speed: Faster with respect to time to first token(TFTT) and tokens per second(tok/s), likely due to using the Cerebras chips. 2. Stamina: More suited effective across longer tasks with auto-compaction. I rarely see any context rot. Also it's very token-efficient without wordy paragraphs. 3. Hiccups: It seems to make sensible implementation choices without pausing at every step. I've rarely had to course correct. 4. Visuals: Diagrams, assets, and image generation feel like part of the same workflow instead of a separate toolchain. Also better integration with more SOTA image generation models. Also Codex+GPT 5.6 Sol made this GIF. PS: I know this is purely anecdotal. When I have time and tokens I can follow up with a more formal comparison.

@Im_IrushiK Exactly this! Opus 5 is such a downgrade. 4.5 to 4.8 were relatively interoperable but 5 is too different and anecdotally worse. I have a hard time believing the benchmarks on this.

We had Kimi K3 recursively self-improve the Cline harness to improve its own performance. 17 hours later, it went from 77.5% to 88.8% on Terminal Bench, and cut run cost from $79 to $49.8.

I was laid off by Amazon AGI today, along with many of my colleagues. My job was deciding which data points matter for pretraining. Turns out I was the one that got filtered out. At 1pm, the reminder for our weekly project sync still went off — for a project that no longer has a team. We could only exchange a smile that needed no words. More seriously: everything I've worked on has been about data, data curation pipeline, data reordering, and data uncertainty. I’m actively looking for new opportunities, including pretraining, trustworthy LLMs, safety, and LLM efficiency — please reach out if you have any openings! To my manager and teammates: thank you for all the help and support you've given me along the way — I learned so much, and I loved the work we did together. It was a great journey, and I know even better ones are ahead for each of you.

Exactly! This is what all these heavily politicized, self-serving AI takes that are emerging against open-source AI completely dismiss. I concede that business dynamics change and that uncertainty is uncomfortable, but if you consider what the history of open-source has been, the benefits are clear for scientific progress and humanity.

Melanie Mitchell @MelMitchell1

2 months ago

In the off chance than anyone cares what I think on this topic: (1) the open-source software movement has been enormously beneficial to socity (2) open-weight (& better, open-data) LLMs will be essential for understanding this technology & for it to be beneficial to society.

27 75 632 97K 38

Taste used to be a byproduct of the reps. Agents took the reps - so if you're junior, you now have to go get the taste on purpose.

Addy Osmani @addyosmani

2 months ago

x.com/i/article/2077…

39 56 537 1.7M 714

GPT‑5.6 is rolling out in Codex. If your best coding sessions live in Claude Code, that shouldn’t stop you from trying it. Introducing openmemory v2. An open-source CLI to port coding sessions across Codex, Claude Code, and OpenCode. Your context should move with your model. github.com/mem0ai/openmem…

OpenAI @OpenAI

2 months ago

Sol, Terra, and Luna, our GPT‑5.6 family of models, are starting to roll out now in ChatGPT, Codex, and the API.

585 1K 12K 4.5M 2K

@addyosmani @aiDotEngineer @swyx @romainhuet @mattyp @lydiahallie @Baconbrix @pbakaus @threepointone @GergelyOrosz Great talk dude. Nice to see someone pushing for putting humans back into the loop.

We took a 30B model and split it in two to write tokens in parallel instead of one at a time. Introducing Nemotron-Labs-TwoTower: a diffusion language model from NVIDIA Research adapted from Nemotron-3-Nano-30B-A3B. Here’s how it works: one half holds the context, the other writes the tokens, with both reusing the pretrained model instead of training a new one from scratch. We found it kept 98.7% of the original model’s quality at 2.42× faster generation.

“Loop engineering” is a hot buzzphrase after mentions of it by Boris Cherny (Claude Code’s creator) and Peter Steinberger (OpenClaw's creator) went viral on social media. Loops are now a key part of how we get AI agents to iterate at length to build software. In this letter, I’d like to share my 3 key loops, shown in the image below, for building 0-to-1 products. These loops guide not just how I build software, but also how I decide what software to build. Agentic coding loop: Given a product specification and optionally a set of evals (that is, a dataset against which to measure performance), we can have an AI agent write code, test its work, and keep iterating until the code is bug-free and meets its specification. This idea of closing the loop took off around the end of last year, and it has been a game changer in enabling coding agents to work longer productively without human intervention. For example, over the weekend, I was building an app for my daughter to practice typing, and my coding agent could easily work for around an hour, using a web browser to check what it had built multiple times before getting back to me, without needing my intervention. The engineering loop executes quickly. Every few minutes, the coding agent might build and test a new version of the software. I hear frequently from developers who are finding new ways to engineer more effective engineering loops. This is an active area of invention! Developer feedback loop: In this loop, a developer examines the current product and steers the coding agent to improve it. Last year, a lot of developers (including me) were acting as the QA (quality assurance) function for our coding agents, manually finding bugs and then asking the agent to fix them. But with coding agents much more able to test their own code, the amount of time we need to spend on this function has decreased significantly. This allows us to make higher-level product decisions, such as what key features to offer, where the UI needs improvement, and so on. The developer-feedback loop operates over time intervals between tens of minutes and hours — that's how frequently a developer might review a product and give feedback. In the case of the typing app, I changed my mind a few times about the visual design, what cat costumes she can unlock as she learns (she loves cats), and the user flow for a grown-up to log in and steer the child's learning experience. When a developer has a clear vision for what to build, it is still a lot of work to translate that vision into a specification for a coding agent to implement. Further, after the developer has seen an implementation, they might update (or perhaps clarify) the spec to steer it toward what they want. If you find that the system repeatedly runs into certain problems, building a set of evals for the agent becomes useful. AI-native teams are increasingly using AI to help shape product direction, for example, automating the gathering and analysis of usage data, summarizing written and verbal customer feedback, or carrying out competitive analysis. However, for pretty much all the products I’m involved in, I see humans as having a significant context advantage over current AI systems — we know a lot more than the AI system about the users and the context the product has to operate in — and thus humans play a critical role. Many people describe this human contribution as “taste,” but I prefer to think of it as humans having a context advantage, since that gives us a clearer path to helping AI systems get better. This also speaks to why this step can’t be automated: So long as the human knows something the AI does not, human-in-the-loop is needed to to inject that knowledge into the system. External feedback loop: This includes a wide range of tactics like asking a few friends for feedback, launching to alpha testers, or putting the code into production with A/B testing. These tactics are usually slow, rarely taking less than hours and sometimes taking days or even weeks. This data informs the developer vision, which in turn continues to drive the detailed product spec, which in turn drives the coding agent. With coding agents speeding up software development, more engineers are starting to play a partial product management role. For many engineers who are growing into this role, the hardest part is shaping the product vision and striking a balance between building (bridging the gap between vision and spec) and getting user feedback to evolve the vision. It is important to do both! I will write more about how to do this in future posts, but for now, I find it encouraging that engineers are playing an expanded role (just as product managers and designers now do more engineering). [Original text: The Batch]

🚨 Forget LIDAR. The Robbyant team just dropped a streaming 3D model that reconstructs scenes live, at ~20 FPS, over long sequences. One single camera. Runs in real time. Open-source. Entirely end-to-end. NO iterative optimization tricks and no post-processing cleanup steps! It outperforms both existing streaming approaches and several offline methods. 100% Free and open-source. Repo, paper and model weights in 🧵↓

This is a new paradigm for interacting with Claude that is significantly more "inline" with all the other human activity org-wide. Once you do all of the under the hood engineering work to make this "just work" (e.g. across tools, integrations, compute environments, memory, security, etc.), Claude basically joins the team in a seamless way - you can talk to it as you would talk to a person and it can help with a very large variety of workloads. Imo this is the 3rd major redesign of LLM UIUX. The first paradigm was that the LLM is a website you go to, the second was that it is an app you download to your computer. This third one is that it is a self-contained, persistent, asynchronous entity with org-wide tools and context, working alongside teams of humans. It really takes a while to wrap your head around it, but it works and it is awesome.

Claude @claudeai

2 months ago

Introducing Claude Tag, a new way for teams to work with Claude. In Slack, Claude joins as a team member with access to the channels and tools you choose. Tag Claude in and delegate tasks to it while you focus on other work.

2K 2K 28K 20.9M 12K

MCP is fixing tool discovery at the wrong layer. The unit of discovery shouldn't be the tool, should be the agent. Instead of retrieving the right tool from a list of 60+, route to a capability-scoped agent. Stop retrieving the right tool. Start routing to the right agent. nishaanthreddy.com/blog/discover-…

// The Efficiency Frontier // Cool paper on context management. As agents reuse the same documents and histories across many turns, the cheapest context strategy is not fixed. This work describes a principled rule for picking one per deployment instead of defaulting to whatever topped a benchmark in isolation. Retrieval and compression methods are almost always benchmarked on accuracy and cost separately, so you never learn when one actually beats another under real load. The Efficiency Frontier models context strategy selection as a single cost-performance problem, with a log-utility term for diminishing returns from extra context and a reuse parameter N that amortizes preprocessing across repeated queries. Sweep N and the optimal strategy changes, exposing crossover regions where retrieval, compression, or full context each wins. On 5,000 HotpotQA instances, deployment-aware selection cuts effective token usage about 25 percent at the same performance, and amortized memory compression runs over 50 percent cheaper than full-context prompting in higher-performance settings. Paper: arxiv.org/abs/2605.23071 Learn to build effective AI agents in our academy: academy.dair.ai

// Adapt the Interface, Not the Model // I am fascinated by the results across my cheap-model-plus-good-harness builds. This new paper also shows good signs of the code-as-agent-harness thesis. The idea is really simple. Do not touch the model. Instead, modify the runtime interface that wraps the frozen LLM. Then convert recurring interaction failures into reusable interventions on the harness side. The paper reports an average relative improvement 88.5% across 7 deterministic environments, 126 model-environment settings, and 18 backbones. A harness learned from one model trajectory generalizes to 17 other backbones. That tells you the harness is capturing environment structure, not model-specific patterns. If you ship agents in production, your harness work is more portable than you might assume. Paper: arxiv.org/abs/2605.22166 Learn to build effective AI agents in our academy: academy.dair.ai