Source: HackerNews AI

AWS旗下 Strands 团队开源了一个专为 AI 代理“快速做决定”打造的模型：Strands Decider 2B。它主打的不是生成长文本，而是在预设选项里完成三件事：分类、工具选择，以及安全性判定——让代理在真正执行任务前先走对路、走稳路。 这款模型规模达到 20 亿参数，能直接在本机 CPU 或 GPU 上低延迟运行。AWS 还同步发布了模型权重、训练资料和训练代码，方便开发者复现与二次训练。 更关键的是它如何接入“代理调用工具”的流程：当智能体准备执行某个动作时，把对话内容和拟调用的工具信息交给 Decider，让它判断该不该继续、走哪条工具路径，甚至识别是否触发安全条件。 例子很直观：用户问天气但没给地点——Decider 可以先阻止代理直接调用天气工具，并促使代理追问地点，避免凭空猜测。 #AIAgents #MachineLearning #OpenSource #AWS #LLM #ToolUse

For now, I tried a few things myself: * * * The short version is: **the determinism claim held up surprisingly well in the external checks I could run** , including a Tesla T4 path and a one-step Qwen3-4B test using the same Q4_K_M base SHA and ToolUse training-data SHA recorded in the published adapter provenance. There was also one useful boundary condition: **holding the binary fixed and changing the host was not enough to break the tiny test, but changing the build profile was enough to change the adapter SHA.** So I think there are really a few different reproducibility contracts hiding under the word “deterministic”, and separating them makes the results easier to interpret. The strongest 4B check I ran was: published base GGUF SHA == probe base GGUF SHA published training SHA == probe training SHA same Runner binary same rank / alpha / lr / ctx / seed semantics same one-step training job CPU path vs Tesla T4-assisted backward The CPU and T4-assisted runs both reported the same one-step loss (`0.676138`) and produced the same adapter SHA-256: 79c5f136486bcad288f0ad2f4f66345dab07db668f0cfe3de9d757bbfdb48e8f The adapter files were byte-for-byte identical. The T4 side was not a silent CPU fallback; Runner reported: train-gpu: Tesla T4 — backward matvec on device The exact base was the bartowski Qwen3-4B Q4_K_M, whose file SHA is: fbe1d5edd4ce802ae3ae7c7e4ab7d09789d697fdac1fc7929f8df4ca3c41bae3 and the regenerated ToolUse training set also matched the SHA recorded in the published adapter provenance. I would still phrase this as an **exact-base/data one-step bridge** , not a reproduction of the full published training run: the published adapter provenance records Runner `0.1.20-alpha` and 316 steps, while my probe used a pinned `0.2.0`-era commit (`8652a7f...`) and deliberately stopped after one update. ### One distinction that may be useful for future reproduction reports After doing these checks, I would probably separate at least three things: 1. Artifact determinism same binary + same model/data/config -> same adapter bytes? 2. Build/toolchain reproducibility same source + independently built binary -> same adapter bytes? 3. Behavioral reproducibility same predictions / probabilities / task behavior? Those did not behave identically in my small tests. With the **same binary** , I could replay the tiny fixture on a second Linux x86_64 host with a different Xeon generation, kernel and glibc version and still get the same adapter SHA. Thread counts `1`, `2`, and `4` also gave the same adapter. But on the same Colab machine, rebuilding the same source with a different ISA/build profile changed the tiny adapter SHA. Interestingly, the resulting parameter difference was extremely small despite the file SHA changing: the adapter cosine was effectively 1 and the maximum F32 difference was only on the order of `1e-7`. So I would treat SHA-256 as a very sensitive **artifact-identity oracle** , but not as a behavioral-distance metric: different SHA does not necessarily mean meaningfully different behavior That distinction may be useful if people start reporting results from ARM, Windows, different compilers, different release binaries, etc. For that reason, a third-party reproduction record might be easier to interpret if it includes something like: Runner commit Runner binary SHA-256 compiler + version build profile / ISA OS + architecture base GGUF SHA-256 training-data SHA-256 seed / rank / alpha / optimizer config adapter SHA-256 loss trajectory rather than treating “same source checkout” and “same executable” as interchangeable. This is also a stronger reproducibility target than the one normally assumed by ML frameworks. For comparison, PyTorch’s reproducibility documentation explicitly does not promise complete reproducibility across releases/platforms or between CPU and GPU execution. For the **BF16 / Q8_0 / Q4_K_M** part, I think there is a relatively cheap next measurement before constructing a substantially harder benchmark. Right now the interesting result is: adapter weights: BF16 ≈ Q8, Q4 somewhat farther away held-out exact-call score: BF16 = Q8 = Q4 = 1.00 That does not necessarily mean the learned functions are equally far apart or equally close. It may just mean the current exact-call metric is saturated. My default next step would be: existing held-out prompts | v teacher-force the gold tool call and compare its token log-prob / NLL | +-- differences visible | -> the precision-dependent path is already | behaviorally visible below top-1 | +-- still almost identical | v compare earlier checkpoints | +-- trajectories separate | -> different learning path, | similar final endpoint | +-- still insensitive -> move to a harder / OOD / near-confusable tool task That seems cheaper than immediately building a large new evaluation suite, and it keeps the claim narrow: first ask whether the existing test hides probability-margin differences; only then ask for a harder task. What I actually tested (click for more details) Where I think this sits relative to nearby GGUF / training work (click for more details) One related branch: merging the adapter back into low-bit weights (click for more details) So my current read is: * **the external determinism check is positive** for the paths I could test; * **binary identity looks like an important part of the reproducibility contract** ; * **adapter SHA and behavioral distance should probably be reported separately** ; * and, for the precision experiment, **gold-sequence probability / NLL looks like the cheapest next discriminator before moving to a harder benchmark**. At least from these checks, I would probably spend the next experiment budget on the BF16/Q8/Q4 behavioral-separation question rather than on another small determinism repetition.

I’ve been working on making Xyntetik Runner train LoRA adapters directly through the same quantized GGUF weights it serves. The important part is that there is no separate FP16 training copy and no second training runtime. The forward pass used for training is the forward pass used for inference. The first reproducible artifact is now live on Hugging Face: huggingface.co ### Joakimpalm-Zen/Qwen3-4B-Runner-ToolUse-Q4_K_M · Hugging Face We’re on a journey to advance and democratize artificial intelligence through open source and open science. For this release I trained a tool-use LoRA directly against a frozen Qwen3-4B Q4_K_M GGUF. On the small held-out eval, exact tool-call accuracy went from 0.69 to 1.00 and right-tool selection from 0.72 to 1.00. More interesting to me is the reproducibility result: same base GGUF * same dataset * same seed * same training config = byte-identical adapter GGUF Two independent runs at 4B scale produced the same SHA256. I also repeated the same training task through BF16, Q8_0 and Q4_K_M copies of the same base model. The BF16 and Q8-trained adapters ended up extremely close in weight space, cosine 0.9998. The Q4-trained adapter followed a measurably different optimization path, cosine 0.9926 against the others, while all three adapters still reached 1.00 on this particular held-out task. So at least in this experiment, training through the deployed quantization and training through the high-precision parent are not numerically the same operation, even though they ended up capability-equivalent on the measured task. The repo contains the adapters, dataset, raw eval JSONs, provenance records, hashes and commands needed to reproduce the runs. A few important limits: This is a narrow synthetic tool-use task with 29 held-out prompts. The training set is small enough that the run is clearly in memorization territory. I’m treating this as a demonstration of the training mechanism and reproducibility properties, not as evidence of general tool-use improvement. I’d be especially interested in people trying the reproducibility claim on hardware I don’t have access to, or rerunning the BF16/Q8/Q4 comparison on a harder task where the optimization-path difference may become behaviorally visible. Runner itself is here: github.com ### GitHub - Joakimpalm-Zen/xyntetik-runner: Single-binary GGUF model runtime in C —... Single-binary GGUF model runtime in C — CPU/CUDA/Metal, OpenAI-compatible. Serves, scores, and trains LoRA directly through the quantized weights it deploys, with byte-reproducible adapters. Tool calls survive the token limit; sparse MoE and schema-constrained decoding included.

Researchers at the University of the Sunshine Coast documented rare dolphin behavior off Australia's Queensland coast, capturing footage of dolphins using shells as tools to extract fish. The study, published this month, reveals two distinct observations: in July 2025, an adult female dolphin was seen lifting gastropod shells out of water to drain fish into her mouth. In October 2025, a mother-calf pair demonstrated the same technique, marking the first evidence that this tool-use behavior is transmitted across generations. These incidents occurred in Hervey Bay, a location previously unassociated with such activity outside Western Australia's Shark Bay. The findings suggest dolphins on the eastern coast independently developed this foraging strategy, though researchers note the behavior may be underreported due to its short duration. The study highlights the complexity of dolphin cognition and the importance of observational research in marine biology, offering insights into animal tool use and cultural transmission in wildlife.

まず、この症状が自分のケースかを切り分けます。今回の issue 群で共通しているのは、claude の実行中に通信まわりが壊れたとき、EPIPE を起点にセッションやプロセスの挙動が崩れる点です。 ひとつは、tooluse のターン中に api.anthropic.com への API 呼び出しが失敗したケースです。報告では、DNS 失敗、HTTP タイムアウト、接続リセットのような一時的なネッ

まず、この症状が自分のケースかを切り分けます。今回の issue 群で共通しているのは、claude の実行中に通信まわりが壊れたとき、EPIPE を起点にセッションやプロセスの挙動が崩れる点です。 ひとつは、tooluse のターン中に api.anthropic.com への API 呼び出しが失敗したケースです。報告では、DNS 失敗、HTTP タイムアウト、接続リセットのような一時的なネッ

まず、この症状が自分のケースかを切り分けます。今回の issue 群で共通しているのは、claude の実行中に通信まわりが壊れたとき、EPIPE を起点にセッションやプロセスの挙動が崩れる点です。 ひとつは、tooluse のターン中に api.anthropic.com への API 呼び出しが失敗したケースです。報告では、DNS 失敗、HTTP タイムアウト、接続リセットのような一時的なネッ

まず、この症状が自分のケースかを切り分けます。今回の issue 群で共通しているのは、claude の実行中に通信まわりが壊れたとき、EPIPE を起点にセッションやプロセスの挙動が崩れる点です。 ひとつは、tooluse のターン中に api.anthropic.com への API 呼び出しが失敗したケースです。報告では、DNS 失敗、HTTP タイムアウト、接続リセットのような一時的なネッ

Large language models are powerful generalists, yet solving deep and complex problems such as those of the Humanity's Last Exam (HLE) remains both conceptually challenging and computationally…

YouTube video by The Canadian Press