quantprobe 1.5.1__tar.gz → 1.6.0__tar.gz

{quantprobe-1.5.1/quantprobe.egg-info → quantprobe-1.6.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: quantprobe
-Version: 1.5.1
+Version: 1.6.0
 Summary: Probe-then-quantize for LLMs: measure a model's fragility curve, plan bit/tier placement by the tiered decode law, and emit ready-to-run llama.cpp recipes.
 Author: Federico Sciuca
 License: MIT
@@ -30,7 +30,7 @@ Dynamic: license-file
 > Pick any model + your machine → predicted tok/s, memory fit, quality cost, and your cheapest next upgrade — from the law below, with your config plotted against every validated measurement.
 
 
-**New here? → [QUICKSTART.md](QUICKSTART.md) gets you running in 60 seconds.** Two tiers: `plan`/`target` and the web calculator need *nothing* installed; `quantize`/`probe`/`run`/`bench`/`dashboard` drive [llama.cpp](https://github.com/ggml-org/llama.cpp/releases) (point quantprobe at it with `--llama-dir`, `QUANTPROBE_LLAMA_DIR`, or `PATH`). Preview any command without llama.cpp using `--dry`. 16 machine presets ship in (`--machine`): GTX 1060 → RTX 5090, Apple `mac-m2/m3/m4-*`, DGX Spark, Epyc — or pass raw specs. Multi-GPU / RAID? Comma lists aggregate: `--vram 24,24 --disk-bw 14,14`. Big-VRAM + disk-streaming rigs get the three-tier expert-cache row (v1.3).
+**New here? → [QUICKSTART.md](QUICKSTART.md) gets you running in 60 seconds — or just `pip install quantprobe && quantprobe auto qwen3-coder --run`.** The calculators (`hw`/`plan`/`target`/`optimize`) need nothing installed; the weight-touching commands drive [llama.cpp](https://github.com/ggml-org/llama.cpp/releases) (point quantprobe at it with `--llama-dir`, `QUANTPROBE_LLAMA_DIR`, or `PATH`; preview anything with `--dry`). 16 machine presets ship in (`--machine`): GTX 1060 → RTX 5090, Apple `mac-m2/m3/m4-*`, DGX Spark, Epyc — or pass raw specs. Multi-GPU / RAID? Comma lists aggregate: `--vram 24,24 --disk-bw 14,14`. Big-VRAM + disk-streaming rigs get the three-tier expert-cache row (v1.3).
 
 
 ---
@@ -67,7 +67,9 @@ Add to that: a **byte-identical control** (two GGUFs the same size, 2.25 ppl apa
 <p align="center"><img src="weights/data/validation_19tok/live_run_20tps.png" width="880" alt="One frame: Task Manager showing 16 GB DDR4-3000 and the GTX 1060 6GB beside llama.cpp chatting Qwen3-30B-A3B live at 20.4 tok/s generation"></p>
 <p align="center"><em>One frame, no cuts: Task Manager (16 GB @ 3000 MT/s, GTX 1060 6 GB, RAM at 91% — the hybrid placement using the whole machine) beside llama.cpp chatting Qwen3-30B-A3B at <b>20.4 tok/s generation</b> — above the pre-registered 19. Raw logs, hardware attestation + GGUF SHA256: <a href="weights/data/validation_19tok/EVIDENCE.txt">EVIDENCE.txt</a>. Third bench run: 19.26 ± 0.45 (series 19.30 → 19.55 → 19.26).</em></p>
 
-**Open pre-registrations** — predictions staked publicly *before* measurement: [colibri v1.1, five falsifiable predictions (2026-07-23)](preregistrations/2026-07-23-colibri-v1.1.md) — dual-SSD scaling, int3 speedup, lattice-vs-scalar, AVX-512 tier-scoping, MTP×MoE antagonism.
+**Open pre-registrations** — predictions staked publicly *before* measurement: [colibri v1.1, five falsifiable predictions (2026-07-23)](preregistrations/2026-07-23-colibri-v1.1.md) — dual-SSD scaling, int3 speedup, lattice-vs-scalar, AVX-512 tier-scoping, MTP×MoE antagonism — plus open bands on the CPU-only context slope and expert-pruning decode ([all verdicts](preregistrations/)).
+
+**Live research track — Law 5 (prefill):** protocol-first and scored same-day in [weights/LAW5_PROTOCOL.md](weights/LAW5_PROTOCOL.md) — 6 staked hits, 4 published misses, and one self-caught retraction corrected in public within the hour, including the measured recipe that cuts an agent's 8k-context turn cost from ~75 s to ~0.5 s. Next up, staked before launch: [Law 6 — speculation economics](preregistrations/2026-07-24-law6-speculation-economics.md), scoring in public the week of 2026-07-27.
 
 
 ## The four placement laws
@@ -96,7 +98,7 @@ quantprobe auto qwen3-coder --tps 15 --run   # empty machine -> optimal quant ch
 
 ```bash
 quantprobe auto qwen3-30b --tps 15                       # ONE command: optimizer picks bits, closest quant fetched, run command printed
-quantprobe auto qwen3-30b --custom                       # THE PRODUCT: probe YOUR model, build its personalized depth-aware GGUF
+quantprobe auto qwen3-30b --custom --run                 # THE PRODUCT: probe YOUR model, build its personalized depth-aware GGUF, launch it
 quantprobe hw                                            # what the law sees on THIS machine (every value source-tagged)
 quantprobe plan     --gguf model.gguf                    # zero-config prediction: placement + tok/s + the launch command
 quantprobe optimize --tps 20                             # CHEAPEST PATH to a target: bits x placement x hardware, Pareto-ranked
@@ -119,14 +121,15 @@ The loop is self-validating: `plan` predicted 17.5 for a file we then measured a
 
 | | |
 |---|---|
-| [QUICKSTART.md](QUICKSTART.md) | 60-second start, three levels; Ollama interop; llama.cpp version notes |
+| [QUICKSTART.md](QUICKSTART.md) | 60-second start, three levels; recipes (own fine-tune, coding agents, hardware buying); Ollama interop |
 | [LAWS.md](LAWS.md) | the four laws — statements, measurements, falsifiable predictions, the general form |
 | [docs/EXAMPLES.md](docs/EXAMPLES.md) | worked examples with real outputs: zero-config, the ×5.4 optimizer A/B, probe walkthrough, troubleshooting |
 | [docs/HARDWARE.md](docs/HARDWARE.md) | the 2016 box: exact specs, measured bandwidths, what the next euro buys |
 | [docs/DEEP-DIVE.md](docs/DEEP-DIVE.md) | what's new vs. what's built-on, parity tables, the 744B-at-home projection, repository map |
 | [preregistrations/](preregistrations/) | every staked prediction with its verdict — hits, the near-miss, and the honest miss |
+| [weights/LAW5_PROTOCOL.md](weights/LAW5_PROTOCOL.md) | the live Law-5 research ledger — conventions, stakes, scores, and the retraction |
 | [papers/arxiv/](papers/arxiv/) | the paper (submission-ready LaTeX) |
-| [CHANGELOG.md](CHANGELOG.md) | v1.0 → v1.4, every release |
+| [CHANGELOG.md](CHANGELOG.md) | v1.0 → v1.5, every release |
 
 ## Honest limitations
 

{quantprobe-1.5.1 → quantprobe-1.6.0}/README.md

@@ -12,7 +12,7 @@
 > Pick any model + your machine → predicted tok/s, memory fit, quality cost, and your cheapest next upgrade — from the law below, with your config plotted against every validated measurement.
 
 
-**New here? → [QUICKSTART.md](QUICKSTART.md) gets you running in 60 seconds.** Two tiers: `plan`/`target` and the web calculator need *nothing* installed; `quantize`/`probe`/`run`/`bench`/`dashboard` drive [llama.cpp](https://github.com/ggml-org/llama.cpp/releases) (point quantprobe at it with `--llama-dir`, `QUANTPROBE_LLAMA_DIR`, or `PATH`). Preview any command without llama.cpp using `--dry`. 16 machine presets ship in (`--machine`): GTX 1060 → RTX 5090, Apple `mac-m2/m3/m4-*`, DGX Spark, Epyc — or pass raw specs. Multi-GPU / RAID? Comma lists aggregate: `--vram 24,24 --disk-bw 14,14`. Big-VRAM + disk-streaming rigs get the three-tier expert-cache row (v1.3).
+**New here? → [QUICKSTART.md](QUICKSTART.md) gets you running in 60 seconds — or just `pip install quantprobe && quantprobe auto qwen3-coder --run`.** The calculators (`hw`/`plan`/`target`/`optimize`) need nothing installed; the weight-touching commands drive [llama.cpp](https://github.com/ggml-org/llama.cpp/releases) (point quantprobe at it with `--llama-dir`, `QUANTPROBE_LLAMA_DIR`, or `PATH`; preview anything with `--dry`). 16 machine presets ship in (`--machine`): GTX 1060 → RTX 5090, Apple `mac-m2/m3/m4-*`, DGX Spark, Epyc — or pass raw specs. Multi-GPU / RAID? Comma lists aggregate: `--vram 24,24 --disk-bw 14,14`. Big-VRAM + disk-streaming rigs get the three-tier expert-cache row (v1.3).
 
 
 ---
@@ -49,7 +49,9 @@ Add to that: a **byte-identical control** (two GGUFs the same size, 2.25 ppl apa
 <p align="center"><img src="weights/data/validation_19tok/live_run_20tps.png" width="880" alt="One frame: Task Manager showing 16 GB DDR4-3000 and the GTX 1060 6GB beside llama.cpp chatting Qwen3-30B-A3B live at 20.4 tok/s generation"></p>
 <p align="center"><em>One frame, no cuts: Task Manager (16 GB @ 3000 MT/s, GTX 1060 6 GB, RAM at 91% — the hybrid placement using the whole machine) beside llama.cpp chatting Qwen3-30B-A3B at <b>20.4 tok/s generation</b> — above the pre-registered 19. Raw logs, hardware attestation + GGUF SHA256: <a href="weights/data/validation_19tok/EVIDENCE.txt">EVIDENCE.txt</a>. Third bench run: 19.26 ± 0.45 (series 19.30 → 19.55 → 19.26).</em></p>
 
-**Open pre-registrations** — predictions staked publicly *before* measurement: [colibri v1.1, five falsifiable predictions (2026-07-23)](preregistrations/2026-07-23-colibri-v1.1.md) — dual-SSD scaling, int3 speedup, lattice-vs-scalar, AVX-512 tier-scoping, MTP×MoE antagonism.
+**Open pre-registrations** — predictions staked publicly *before* measurement: [colibri v1.1, five falsifiable predictions (2026-07-23)](preregistrations/2026-07-23-colibri-v1.1.md) — dual-SSD scaling, int3 speedup, lattice-vs-scalar, AVX-512 tier-scoping, MTP×MoE antagonism — plus open bands on the CPU-only context slope and expert-pruning decode ([all verdicts](preregistrations/)).
+
+**Live research track — Law 5 (prefill):** protocol-first and scored same-day in [weights/LAW5_PROTOCOL.md](weights/LAW5_PROTOCOL.md) — 6 staked hits, 4 published misses, and one self-caught retraction corrected in public within the hour, including the measured recipe that cuts an agent's 8k-context turn cost from ~75 s to ~0.5 s. Next up, staked before launch: [Law 6 — speculation economics](preregistrations/2026-07-24-law6-speculation-economics.md), scoring in public the week of 2026-07-27.
 
 
 ## The four placement laws
@@ -78,7 +80,7 @@ quantprobe auto qwen3-coder --tps 15 --run   # empty machine -> optimal quant ch
 
 ```bash
 quantprobe auto qwen3-30b --tps 15                       # ONE command: optimizer picks bits, closest quant fetched, run command printed
-quantprobe auto qwen3-30b --custom                       # THE PRODUCT: probe YOUR model, build its personalized depth-aware GGUF
+quantprobe auto qwen3-30b --custom --run                 # THE PRODUCT: probe YOUR model, build its personalized depth-aware GGUF, launch it
 quantprobe hw                                            # what the law sees on THIS machine (every value source-tagged)
 quantprobe plan     --gguf model.gguf                    # zero-config prediction: placement + tok/s + the launch command
 quantprobe optimize --tps 20                             # CHEAPEST PATH to a target: bits x placement x hardware, Pareto-ranked
@@ -101,14 +103,15 @@ The loop is self-validating: `plan` predicted 17.5 for a file we then measured a
 
 | | |
 |---|---|
-| [QUICKSTART.md](QUICKSTART.md) | 60-second start, three levels; Ollama interop; llama.cpp version notes |
+| [QUICKSTART.md](QUICKSTART.md) | 60-second start, three levels; recipes (own fine-tune, coding agents, hardware buying); Ollama interop |
 | [LAWS.md](LAWS.md) | the four laws — statements, measurements, falsifiable predictions, the general form |
 | [docs/EXAMPLES.md](docs/EXAMPLES.md) | worked examples with real outputs: zero-config, the ×5.4 optimizer A/B, probe walkthrough, troubleshooting |
 | [docs/HARDWARE.md](docs/HARDWARE.md) | the 2016 box: exact specs, measured bandwidths, what the next euro buys |
 | [docs/DEEP-DIVE.md](docs/DEEP-DIVE.md) | what's new vs. what's built-on, parity tables, the 744B-at-home projection, repository map |
 | [preregistrations/](preregistrations/) | every staked prediction with its verdict — hits, the near-miss, and the honest miss |
+| [weights/LAW5_PROTOCOL.md](weights/LAW5_PROTOCOL.md) | the live Law-5 research ledger — conventions, stakes, scores, and the retraction |
 | [papers/arxiv/](papers/arxiv/) | the paper (submission-ready LaTeX) |
-| [CHANGELOG.md](CHANGELOG.md) | v1.0 → v1.4, every release |
+| [CHANGELOG.md](CHANGELOG.md) | v1.0 → v1.5, every release |
 
 ## Honest limitations
 

{quantprobe-1.5.1 → quantprobe-1.6.0}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "quantprobe"
-version = "1.5.1"
+version = "1.6.0"
 description = "Probe-then-quantize for LLMs: measure a model's fragility curve, plan bit/tier placement by the tiered decode law, and emit ready-to-run llama.cpp recipes."
 readme = "README.md"
 license = { text = "MIT" }

{quantprobe-1.5.1 → quantprobe-1.6.0}/quantprobe/__init__.py

@@ -8,4 +8,4 @@ Three commands, all grounded in measured laws (see LAWS.md in the repo):
 Every constant in `plan` is fitted from measurements published in the repo, including two
 pre-registered hardware predictions confirmed within 8%.
 """
-__version__ = "1.5.1"
+__version__ = "1.6.0"

{quantprobe-1.5.1 → quantprobe-1.6.0}/quantprobe/auto.py

@@ -68,7 +68,36 @@ def pick_source(files, t):
     return hi[0] if hi else max(cands)
 
 
+def _wizard(a):
+    """No model given: detect, ask two questions, go. The full pipeline with zero flags."""
+    print("quantprobe auto - interactive (no flags needed; `quantprobe auto --help` for the full list)")
+    try:
+        from . import detect as detmod
+        d, _ = detmod.detect()
+        print(f"\n  this machine [auto-detected]: {d['vram']:g} GB VRAM @ {d['vram_bw']:g} GB/s | "
+              f"{d['ram']:g} GB RAM @ {d['ram_bw']:g} GB/s | disk {d['disk_bw']:g} GB/s")
+    except Exception:
+        pass
+    try:
+        m = input(f"\n  model - preset ({', '.join(MODEL_REPOS)}) or HF GGUF repo id [qwen3-30b]: ").strip() or "qwen3-30b"
+        print("\n  [1] best standard quant for this machine, ready to run  (skips quantization)")
+        print("  [2] full custom: probe YOUR model's fragile layers (~30-60 min), build its personalized GGUF")
+        print("  [3] hit a speed target (asks for tok/s)")
+        c = input("  choice [1]: ").strip() or "1"
+        if c == "2":
+            a.custom = True
+        elif c == "3":
+            a.tps = float(input("  target tok/s: ").strip())
+        r = input("  launch chat when ready? [Y/n]: ").strip().lower()
+        a.run = r != "n"
+    except EOFError:
+        raise SystemExit("no model given and no terminal to ask. Pass one: quantprobe auto qwen3-30b [--custom] [--run]")
+    a.target = m
+
+
 def run(a):
+    if a.target is None:
+        _wizard(a)
     target = a.target
     if target in MODEL_REPOS:
         repo, t, ac, ne, moe = MODEL_REPOS[target]
@@ -111,6 +140,12 @@ def run(a):
     scored.sort()
     _, _, path, size, bits, best = scored[0]
 
+    if getattr(a, "custom", False) and want_bits >= 3.5 and not getattr(a, "force_custom", False):
+        print(f"\n[quantprobe auto --custom] this machine doesn't need the surgery: the optimizer")
+        print(f"  wants ~{want_bits:g}-bit here, and at >=3.5 bits standard community quants match the")
+        print(f"  depth-aware recipe on quality - the fragile-band fix only pays below ~3 bits (Laws 1-2).")
+        print(f"  Fetching the optimal standard quant instead. Build anyway: --force-custom.")
+        a.custom = False
     if getattr(a, "custom", False):
         src = pick_source(files, t)
         if not src:

{quantprobe-1.5.1 → quantprobe-1.6.0}/quantprobe/cli.py

@@ -92,14 +92,16 @@ def main():
     hwargs(o)
 
     au = sub.add_parser("auto", help="ONE command: model in, running setup out - optimizer picks the bits, the closest quant is fetched, run command printed")
-    au.add_argument("target", help="model preset (qwen3-30b, qwen3-coder, glm-air, laguna-s, gemma-12b, mistral-7b) or a HF GGUF repo id")
+    au.add_argument("target", nargs="?", default=None,
+                    help="model preset (qwen3-30b, qwen3-coder, glm-air, laguna-s, gemma-12b, mistral-7b) or a HF GGUF repo id; omit it and quantprobe ASKS (interactive)")
     au.add_argument("--dir", default="./models", help="download directory (default ./models)")
     au.add_argument("--run", action="store_true", help="launch chat immediately after the download")
     hwargs(au)
     au.add_argument("--tps", type=float, default=None, help="target tok/s for the optimizer")
     au.add_argument("--max-quality", type=float, default=None)
     au.add_argument("--allow-prune", action="store_true"); au.add_argument("--any-runtime", action="store_true")
-    au.add_argument("--custom", action="store_true", help="THE PRODUCT: fetch a high-precision source, probe YOUR model (~30-60 min), build a depth-aware GGUF personalized to it")
+    au.add_argument("--custom", action="store_true", help="THE PRODUCT: fetch a high-precision source, probe YOUR model (~30-60 min), build a depth-aware GGUF personalized to it (machine-gated: skipped with an explanation when this box doesn't need sub-3-bit surgery)")
+    au.add_argument("--force-custom", action="store_true", help="build the depth-aware GGUF even when the optimizer says this machine doesn't need it")
     au.add_argument("--serve", action="store_true", help="with --run: llama-server instead of chat")
     au.add_argument("--extra", default=None)
 

{quantprobe-1.5.1 → quantprobe-1.6.0}/quantprobe/optimize.py

@@ -35,6 +35,7 @@ HW_DELTAS = [
 
 
 def resolve(a):
+    planmod.check_presets(a)
     """Model + machine resolution, same semantics as plan.run (autospec + autodetect included)."""
     from . import spec as specmod
     specmod.apply(a, quiet=True)

{quantprobe-1.5.1 → quantprobe-1.6.0}/quantprobe/plan.py

@@ -144,9 +144,27 @@ def qual_of(moe, bits):
     return QUAL[moe][min(keys, key=lambda k: abs(k - bits))]
 
 
+def check_presets(args):
+    """Refuse unknown preset names LOUDLY instead of silently falling back to defaults."""
+    mdl = getattr(args, "model", None)
+    if mdl and mdl not in MODELS and not getattr(args, "total", None):
+        raise SystemExit(
+            "unknown --model '%s' (and no --total to describe it).\n"
+            "  presets: %s\n"
+            "  or describe it:      --total <B> --active <B> [--always-active <B>]\n"
+            "  or point at a file:  --gguf model.gguf   (exact spec read from the GGUF)" % (mdl, ", ".join(sorted(MODELS))))
+    mac = getattr(args, "machine", None)
+    if mac and mac not in MACHINES:
+        raise SystemExit(
+            "unknown --machine '%s'.\n"
+            "  presets: %s\n"
+            "  or pass flags: --vram/--vram-bw/--ram/--ram-bw/--disk-bw   (no flags = auto-detect this box)" % (mac, ", ".join(sorted(MACHINES))))
+
+
 def run(args):
     from . import spec as specmod
     specmod.apply(args)
+    check_presets(args)
     if getattr(args, "bits", None) is None:
         args.bits = 2.5
     m = dict(MODELS[args.model]) if args.model in MODELS else {}

{quantprobe-1.5.1 → quantprobe-1.6.0}/quantprobe/target.py

@@ -52,6 +52,8 @@ def feasible(a, tps_target):
 
 
 def run(a):
+    from . import plan as planmod
+    planmod.check_presets(a)
     hw, vc, vb, rc, rb, db, geta, gl = hw_of(a)
     rows = feasible(a, a.tps)
     print(f"\nquantprobe target - smartest model meeting >= {a.tps:g} tok/s "

{quantprobe-1.5.1 → quantprobe-1.6.0/quantprobe.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: quantprobe
-Version: 1.5.1
+Version: 1.6.0
 Summary: Probe-then-quantize for LLMs: measure a model's fragility curve, plan bit/tier placement by the tiered decode law, and emit ready-to-run llama.cpp recipes.
 Author: Federico Sciuca
 License: MIT
@@ -30,7 +30,7 @@ Dynamic: license-file
 > Pick any model + your machine → predicted tok/s, memory fit, quality cost, and your cheapest next upgrade — from the law below, with your config plotted against every validated measurement.
 
 
-**New here? → [QUICKSTART.md](QUICKSTART.md) gets you running in 60 seconds.** Two tiers: `plan`/`target` and the web calculator need *nothing* installed; `quantize`/`probe`/`run`/`bench`/`dashboard` drive [llama.cpp](https://github.com/ggml-org/llama.cpp/releases) (point quantprobe at it with `--llama-dir`, `QUANTPROBE_LLAMA_DIR`, or `PATH`). Preview any command without llama.cpp using `--dry`. 16 machine presets ship in (`--machine`): GTX 1060 → RTX 5090, Apple `mac-m2/m3/m4-*`, DGX Spark, Epyc — or pass raw specs. Multi-GPU / RAID? Comma lists aggregate: `--vram 24,24 --disk-bw 14,14`. Big-VRAM + disk-streaming rigs get the three-tier expert-cache row (v1.3).
+**New here? → [QUICKSTART.md](QUICKSTART.md) gets you running in 60 seconds — or just `pip install quantprobe && quantprobe auto qwen3-coder --run`.** The calculators (`hw`/`plan`/`target`/`optimize`) need nothing installed; the weight-touching commands drive [llama.cpp](https://github.com/ggml-org/llama.cpp/releases) (point quantprobe at it with `--llama-dir`, `QUANTPROBE_LLAMA_DIR`, or `PATH`; preview anything with `--dry`). 16 machine presets ship in (`--machine`): GTX 1060 → RTX 5090, Apple `mac-m2/m3/m4-*`, DGX Spark, Epyc — or pass raw specs. Multi-GPU / RAID? Comma lists aggregate: `--vram 24,24 --disk-bw 14,14`. Big-VRAM + disk-streaming rigs get the three-tier expert-cache row (v1.3).
 
 
 ---
@@ -67,7 +67,9 @@ Add to that: a **byte-identical control** (two GGUFs the same size, 2.25 ppl apa
 <p align="center"><img src="weights/data/validation_19tok/live_run_20tps.png" width="880" alt="One frame: Task Manager showing 16 GB DDR4-3000 and the GTX 1060 6GB beside llama.cpp chatting Qwen3-30B-A3B live at 20.4 tok/s generation"></p>
 <p align="center"><em>One frame, no cuts: Task Manager (16 GB @ 3000 MT/s, GTX 1060 6 GB, RAM at 91% — the hybrid placement using the whole machine) beside llama.cpp chatting Qwen3-30B-A3B at <b>20.4 tok/s generation</b> — above the pre-registered 19. Raw logs, hardware attestation + GGUF SHA256: <a href="weights/data/validation_19tok/EVIDENCE.txt">EVIDENCE.txt</a>. Third bench run: 19.26 ± 0.45 (series 19.30 → 19.55 → 19.26).</em></p>
 
-**Open pre-registrations** — predictions staked publicly *before* measurement: [colibri v1.1, five falsifiable predictions (2026-07-23)](preregistrations/2026-07-23-colibri-v1.1.md) — dual-SSD scaling, int3 speedup, lattice-vs-scalar, AVX-512 tier-scoping, MTP×MoE antagonism.
+**Open pre-registrations** — predictions staked publicly *before* measurement: [colibri v1.1, five falsifiable predictions (2026-07-23)](preregistrations/2026-07-23-colibri-v1.1.md) — dual-SSD scaling, int3 speedup, lattice-vs-scalar, AVX-512 tier-scoping, MTP×MoE antagonism — plus open bands on the CPU-only context slope and expert-pruning decode ([all verdicts](preregistrations/)).
+
+**Live research track — Law 5 (prefill):** protocol-first and scored same-day in [weights/LAW5_PROTOCOL.md](weights/LAW5_PROTOCOL.md) — 6 staked hits, 4 published misses, and one self-caught retraction corrected in public within the hour, including the measured recipe that cuts an agent's 8k-context turn cost from ~75 s to ~0.5 s. Next up, staked before launch: [Law 6 — speculation economics](preregistrations/2026-07-24-law6-speculation-economics.md), scoring in public the week of 2026-07-27.
 
 
 ## The four placement laws
@@ -96,7 +98,7 @@ quantprobe auto qwen3-coder --tps 15 --run   # empty machine -> optimal quant ch
 
 ```bash
 quantprobe auto qwen3-30b --tps 15                       # ONE command: optimizer picks bits, closest quant fetched, run command printed
-quantprobe auto qwen3-30b --custom                       # THE PRODUCT: probe YOUR model, build its personalized depth-aware GGUF
+quantprobe auto qwen3-30b --custom --run                 # THE PRODUCT: probe YOUR model, build its personalized depth-aware GGUF, launch it
 quantprobe hw                                            # what the law sees on THIS machine (every value source-tagged)
 quantprobe plan     --gguf model.gguf                    # zero-config prediction: placement + tok/s + the launch command
 quantprobe optimize --tps 20                             # CHEAPEST PATH to a target: bits x placement x hardware, Pareto-ranked
@@ -119,14 +121,15 @@ The loop is self-validating: `plan` predicted 17.5 for a file we then measured a
 
 | | |
 |---|---|
-| [QUICKSTART.md](QUICKSTART.md) | 60-second start, three levels; Ollama interop; llama.cpp version notes |
+| [QUICKSTART.md](QUICKSTART.md) | 60-second start, three levels; recipes (own fine-tune, coding agents, hardware buying); Ollama interop |
 | [LAWS.md](LAWS.md) | the four laws — statements, measurements, falsifiable predictions, the general form |
 | [docs/EXAMPLES.md](docs/EXAMPLES.md) | worked examples with real outputs: zero-config, the ×5.4 optimizer A/B, probe walkthrough, troubleshooting |
 | [docs/HARDWARE.md](docs/HARDWARE.md) | the 2016 box: exact specs, measured bandwidths, what the next euro buys |
 | [docs/DEEP-DIVE.md](docs/DEEP-DIVE.md) | what's new vs. what's built-on, parity tables, the 744B-at-home projection, repository map |
 | [preregistrations/](preregistrations/) | every staked prediction with its verdict — hits, the near-miss, and the honest miss |
+| [weights/LAW5_PROTOCOL.md](weights/LAW5_PROTOCOL.md) | the live Law-5 research ledger — conventions, stakes, scores, and the retraction |
 | [papers/arxiv/](papers/arxiv/) | the paper (submission-ready LaTeX) |
-| [CHANGELOG.md](CHANGELOG.md) | v1.0 → v1.4, every release |
+| [CHANGELOG.md](CHANGELOG.md) | v1.0 → v1.5, every release |
 
 ## Honest limitations