Dataset Viewer
Auto-converted to Parquet Duplicate
audio
audioduration (s)
2.07
7.58
id
stringlengths
23
23
text
stringlengths
34
104
text_he
stringlengths
34
104
duration_s
float64
2.07
7.58
categories
listlengths
2
4
target_terms
listlengths
1
1
prompt_id
stringlengths
8
12
prompt_source
stringclasses
2 values
prompt_origin
stringclasses
2 values
mode
stringclasses
1 value
edited
bool
1 class
device
stringclasses
2 values
microphone
stringclasses
5 values
environment
stringclasses
2 values
conditions
listlengths
0
0
noise_level
stringclasses
0 values
background_speech_languages
listlengths
0
0
noise_flags
listlengths
0
0
clipped
bool
1 class
notes
stringclasses
0 values
hebrew_pronunciation
stringclasses
0 values
variant_groups
listlengths
0
0
variant_spelling
stringclasses
0 values
task
stringclasses
2 values
priority_word
stringclasses
64 values
spoken_text
stringlengths
34
104
⌀
aside_text
stringclasses
0 values
aside_position
stringclasses
0 values
verbatim_text
stringclasses
0 values
prompt_categories
listlengths
source_prompt_id
stringclasses
0 values
prompt_kind
stringclasses
0 values
prompt_note
stringclasses
0 values
term_language
stringclasses
0 values
paragraph_text
stringclasses
0 values
model_hypothesis
stringclasses
0 values
asr_model
stringclasses
0 values
speaker
stringclasses
1 value
speaker_id
stringdate
8992-01-01 00:00:00
8992-01-01 00:00:00
contributor
stringclasses
1 value
speaker_gender
stringclasses
1 value
sample_rate
int64
16k
48k
bit_depth
int64
16
24
audio_format
stringclasses
2 values
native_rate
int64
48k
48k
⌀
capture_mode
stringclasses
1 value
tts_eligible
bool
2 classes
tts_split
stringclasses
0 values
transcript_status
stringclasses
1 value
prompt_set
stringclasses
0 values
topic
stringclasses
0 values
style
stringclasses
0 values
style_instruction
stringclasses
0 values
written_form_uncertain
bool
0 classes
session_id
stringclasses
73 values
room_tone_id
stringclasses
0 values
room
stringclasses
0 values
location
stringclasses
0 values
location_name
stringclasses
0 values
location_note
stringclasses
0 values
latitude
float64
longitude
float64
location_accuracy_m
float64
lead_trim_s
float64
0
5.46
⌀
peak_dbfs
float64
-19.8
-1
⌀
floor_dbfs
float64
-79.1
-21.2
⌀
speech_dbfs
float64
-36.3
-7.9
⌀
snr_db
float64
9.8
69.3
⌀
clipped_samples
int64
0
0
⌀
edge_head_dbfs
float64
-120
-1.8
⌀
edge_tail_dbfs
float64
-72.2
-1
⌀
browser_processing
bool
1 class
recorded_at
stringlengths
24
24
text_full
stringclasses
0 values
fallback_ok
bool
0 classes
sentence_id
stringlengths
12
12
pronunciation
stringclasses
1 value
extra
stringclasses
12 values
20261021T120639Z_jxq5mi
Switch the output style to something shorter, I don't need the explanations.
Switch the output style to something shorter, I don't need the explanations.
3.57
[ "tech", "ai_interactions" ]
[ "output style" ]
p_01ad4653eb
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
mixed
null
null
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
16,000
16
wav
null
null
null
null
script
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
2026-10-21T12:06:39.644Z
null
null
p_01ad4653eb
[]
null
20261021T122606Z_bi9i90
Send the Explore agent to find where the bucket name is set.
Send the Explore agent to find where the bucket name is set.
3.89
[ "tech", "ai_interactions" ]
[ "Explore agent" ]
p_e5090d2901
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
mixed
null
null
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
16,000
16
wav
null
null
null
null
script
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
2026-10-21T12:26:06.038Z
null
null
p_e5090d2901
[]
null
20261021T123220Z_e8fhji
Aider made a commit for every small change, squash them before the PR.
Aider made a commit for every small change, squash them before the PR.
4.758
[ "tech", "ai_interactions" ]
[ "Aider" ]
p_565c3ab0d4
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
mixed
null
null
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
16,000
16
wav
null
null
null
null
script
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
2026-10-21T12:32:20.839Z
null
null
p_565c3ab0d4
[]
null
20261021T123227Z_fr61nh
Please amend the last commit, you left the debug line in.
Please amend the last commit, you left the debug line in.
4.346
[ "tech", "ai_interactions" ]
[ "amend" ]
p_1a15cc3a40
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
null
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
16,000
16
wav
null
null
null
null
script
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
2026-10-21T12:32:27.063Z
null
null
p_1a15cc3a40
[]
null
20261021T130621Z_4q5xwf
The hooks are slowing every tool call down, so time each one and tell me the worst.
The hooks are slowing every tool call down, so time each one and tell me the worst.
4.802
[ "tech", "ai_interactions" ]
[ "hooks" ]
p_7722d5dbde
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
null
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
16,000
16
wav
null
null
null
null
script
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
2026-10-21T13:06:21.785Z
null
null
p_7722d5dbde
[]
null
20261022T123215Z_puwg5y
Is there already a GitHub issue upstream about this crash?
Is there already a GitHub issue upstream about this crash?
3.453
[ "tech", "ai_interactions" ]
[ "GitHub issue" ]
p_55af1611a3
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
Is there already a GitHub issue upstream about this crash?
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261022T123205Z_11bhy4
null
null
null
null
null
null
null
null
0.84
-5.2
-43.8
-16.9
27
0
-34.6
-33.5
false
2026-10-22T12:32:15.433Z
null
null
p_55af1611a3
[]
null
20261022T124848Z_8cu2mr
Keep a little stash of chocolate in the drawer for the late nights.
Keep a little stash of chocolate in the drawer for the late nights.
3.343
[ "tech", "ai_interactions" ]
[ "stash" ]
p_d0c2ea09da
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
Keep a little stash of chocolate in the drawer for the late nights.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261022T124320Z_xj92ob
null
null
null
null
null
null
null
null
1.12
-5.8
-52.9
-19.5
33.4
0
-45.9
-28.9
false
2026-10-22T12:48:48.975Z
null
null
p_d0c2ea09da
[]
null
20261023T092040Z_i43oiz
Convert the PyTorch model to ONNX before deploying it.
Convert the PyTorch model to ONNX before deploying it.
3.678
[ "tech", "ai_interactions", "target_word" ]
[ "ONNX" ]
pw_1eynl4g
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
ONNX
Convert the PyTorch model to ONNX before deploying it.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261023T091323Z_3ul7bp
null
null
null
null
null
null
null
null
1.06
-5
-51.4
-20.1
31.3
0
-39.5
-39.6
false
2026-10-23T09:20:40.661Z
null
null
p_eabafe79d0
[]
null
20261023T092319Z_3ra2nb
Can you export the trained network to ONNX format?
Can you export the trained network to ONNX format?
3.671
[ "tech", "ai_interactions", "target_word" ]
[ "ONNX" ]
pw_i9xjnq
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
ONNX
Can you export the trained network to ONNX format?
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261023T091323Z_3ul7bp
null
null
null
null
null
null
null
null
0.72
-9.3
-51.3
-21.9
29.4
0
-39.1
-38.3
false
2026-10-23T09:23:19.747Z
null
null
p_08ee6a60f1
[]
null
20261023T092531Z_oulkvk
Please pull the latest Nemotron checkpoint and run the streaming ASR test.
Please pull the latest Nemotron checkpoint and run the streaming ASR test.
4.827
[ "tech", "ai_interactions", "target_word" ]
[ "Nemotron" ]
pw_jv3zyr
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
Nemotron
Please pull the latest Nemotron checkpoint and run the streaming ASR test.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261023T091323Z_3ul7bp
null
null
null
null
null
null
null
null
1.14
-3.8
-52.1
-19.4
32.7
0
-39.4
-40.6
false
2026-10-23T09:25:31.247Z
null
null
p_a8863245a4
[]
null
20261023T092544Z_j0seso
The ONNX runtime is faster than the native one.
The ONNX runtime is faster than the native one.
3.394
[ "tech", "ai_interactions", "target_word" ]
[ "ONNX" ]
pw_1o8sr3p
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
ONNX
The ONNX runtime is faster than the native one.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261023T091323Z_3ul7bp
null
null
null
null
null
null
null
null
1.24
-6.3
-53.1
-19.6
33.4
0
-40.4
-42.3
false
2026-10-23T09:25:44.316Z
null
null
p_728b103674
[]
null
20261023T105152Z_usgeu8
Why is the Nemotron output drifting after ten seconds of audio?
Why is the Nemotron output drifting after ten seconds of audio?
4.967
[ "tech", "ai_interactions", "target_word" ]
[ "Nemotron" ]
pw_1ujkooo
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
Nemotron
Why is the Nemotron output drifting after ten seconds of audio?
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261023T104353Z_1jzzlh
null
null
null
null
null
null
null
null
1.48
-5.5
-55.9
-17.9
38
0
-45.4
-41.4
false
2026-10-23T10:51:52.831Z
null
null
p_b0cb4c4175
[]
null
20261023T105334Z_njzend
I need the ONNX file for the edge device.
I need the ONNX file for the edge device.
3.342
[ "tech", "ai_interactions", "target_word" ]
[ "ONNX" ]
pw_cw4zj1
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
ONNX
I need the ONNX file for the edge device.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261023T104353Z_1jzzlh
null
null
null
null
null
null
null
null
1.18
-4.2
-55.7
-17.6
38.2
0
-38.3
-46.2
false
2026-10-23T10:53:34.397Z
null
null
p_5da28d1c94
[]
null
20261023T105357Z_9u8oe3
Switch to plan mode and tell me how you'd restructure the training script.
Switch to plan mode and tell me how you'd restructure the training script.
4.45
[ "tech", "ai_interactions" ]
[ "plan mode" ]
p_6e8ec02845
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
Switch to plan mode and tell me how you'd restructure the training script.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261023T104353Z_1jzzlh
null
null
null
null
null
null
null
null
1.08
-2.3
-50.1
-16.3
33.8
0
-39.2
-38.7
false
2026-10-23T10:53:57.766Z
null
null
p_6e8ec02845
[]
null
20261023T111444Z_budqdh
Commit the config change before you swap in the Nemotron weights.
Commit the config change before you swap in the Nemotron weights.
3.641
[ "tech", "ai_interactions", "target_word" ]
[ "Nemotron" ]
pw_136pkkx
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
Nemotron
Commit the config change before you swap in the Nemotron weights.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261023T104353Z_1jzzlh
null
null
null
null
null
null
null
null
1.26
-9.4
-39.4
-21.3
18.1
0
-30.1
-28.9
false
2026-10-23T11:14:44.967Z
null
null
p_db0e62be07
[]
null
20261023T113503Z_dwpflz
Is there anything left in the stash from last week?
Is there anything left in the stash from last week?
3.035
[ "tech", "ai_interactions" ]
[ "stash" ]
p_d5c973e3ad
generated:claude (Claude Code subagent)
generated
bank
false
ThinkPad E15 Gen 3
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
Is there anything left in the stash from last week?
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261023T113411Z_lx0rx0
null
null
null
null
null
null
null
null
0.42
-2.8
-35.7
-17.9
17.8
0
-24.2
-23.6
false
2026-10-23T11:35:03.003Z
null
null
p_d5c973e3ad
[]
null
20261023T114015Z_d86c8g
Please use the ONNX version of the model for inference.
Please use the ONNX version of the model for inference.
3.701
[ "tech", "ai_interactions", "target_word" ]
[ "ONNX" ]
pw_m6aqbx
on-demand
manual
bank
false
ThinkPad E15 Gen 3
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
ONNX
Please use the ONNX version of the model for inference.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261023T113948Z_wzyrwf
null
null
null
null
null
null
null
null
0
-14.7
-43.3
-28.2
15.1
0
-31.5
-31.4
false
2026-10-23T11:40:15.072Z
null
null
p_d899ac8d0d
[]
null
20261023T120958Z_d2wtel
Pull the model weights from Hugging Face and cache them locally before you start training.
Pull the model weights from Hugging Face and cache them locally before you start training.
4.615
[ "tech", "ai_interactions", "target_word" ]
[ "Hugging Face" ]
pw_pl943k
on-demand
manual
bank
false
ThinkPad E15 Gen 3
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
null
Pull the model weights from Hugging Face and cache them locally before you start training.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261023T120245Z_do4ncu
null
null
null
null
null
null
null
null
0.6
-13.4
-43.5
-27.5
16
0
-33.1
-33.8
false
2026-10-23T12:09:58.693Z
null
null
p_1a8c08bc15
[]
null
20261023T121057Z_mxdwv8
I want you to benchmark Nemotron against our current model on the test set.
I want you to benchmark Nemotron against our current model on the test set.
4.758
[ "tech", "ai_interactions", "target_word" ]
[ "Nemotron" ]
pw_1hgqp9c
on-demand
manual
bank
false
ThinkPad E15 Gen 3
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
Nemotron
I want you to benchmark Nemotron against our current model on the test set.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261023T120245Z_do4ncu
null
null
null
null
null
null
null
null
0.66
-16.5
-43.7
-28.2
15.5
0
-34.1
-33.1
false
2026-10-23T12:10:57.652Z
null
null
p_2f8f729276
[]
null
20261023T121222Z_zcetzk
Why is the ONNX conversion failing on this layer?
Why is the ONNX conversion failing on this layer?
3.935
[ "tech", "ai_interactions", "target_word" ]
[ "ONNX" ]
pw_5lr5wc
on-demand
manual
bank
false
ThinkPad E15 Gen 3
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
ONNX
Why is the ONNX conversion failing on this layer?
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261023T120245Z_do4ncu
null
null
null
null
null
null
null
null
0
-12.8
-44
-27.9
16.1
0
-32.4
-33.2
false
2026-10-23T12:12:22.347Z
null
null
p_ab58221ff7
[]
null
20261023T123522Z_e4ccgs
Please implement LoRA in that training script.
Please implement LoRA in that training script.
7.582
[ "tech", "ai_interactions", "target_word" ]
[ "LoRA" ]
pw_wm6smv
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
LoRA
Please implement LoRA in that training script.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261023T123454Z_4imju3
null
null
null
null
null
null
null
null
1.06
-7.3
-44.8
-25.2
19.6
0
-33.3
-32.2
false
2026-10-23T12:35:22.949Z
null
null
p_ced0dfadad
[]
null
20261024T090328Z_ubkgik
Run the abliteration script we discussed, but keep the safety layers intact for now.
Run the abliteration script we discussed, but keep the safety layers intact for now.
4.283
[ "tech", "ai_interactions", "target_word" ]
[ "abliteration" ]
pw_8c4cdc
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
abliteration
Run the abliteration script we discussed, but keep the safety layers intact for now.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261024T090156Z_7sxyfx
null
null
null
null
null
null
null
null
0.9
-5.8
-51.5
-22.8
28.7
0
-39.4
-43.3
false
2026-10-24T09:03:28.296Z
null
null
p_27b366f00e
[]
null
20261024T090352Z_yzjsnx
Run the evals again after you swap in the smaller model.
Run the evals again after you swap in the smaller model.
3.365
[ "tech", "ai_interactions", "target_word" ]
[ "evals" ]
pw_132wq53
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
evals
Run the evals again after you swap in the smaller model.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261024T090156Z_7sxyfx
null
null
null
null
null
null
null
null
0.8
-3.3
-51.9
-21.5
30.4
0
-39.5
-40.8
false
2026-10-24T09:03:52.906Z
null
null
p_a139628806
[]
null
20261024T134416Z_msnpzj
Switch the summarizer over to Sonnet and rerun the failing test.
Switch the summarizer over to Sonnet and rerun the failing test.
3.638
[ "tech", "ai_interactions", "target_word" ]
[ "Sonnet" ]
pw_o84r2h
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
Sonnet
Switch the summarizer over to Sonnet and rerun the failing test.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261024T134105Z_no661v
null
null
null
null
null
null
null
null
0.62
-6.1
-50.8
-17.6
33.2
0
-38.2
-41.6
false
2026-10-24T13:44:16.988Z
null
null
p_c8f5a715d5
[]
null
20261024T181254Z_dz2ujq
Please rebuild the docs site and tell me if Muse Spark finished cleanly.
Please rebuild the docs site and tell me if Muse Spark finished cleanly.
3.89
[ "tech", "ai_interactions", "target_word" ]
[ "Muse Spark" ]
pw_1mzn837
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
Muse Spark
Please rebuild the docs site and tell me if Muse Spark finished cleanly.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261024T134105Z_no661v
null
null
null
null
null
null
null
null
1.24
-3.7
-52.2
-18
34.2
0
-42
-40.9
false
2026-10-24T18:12:54.338Z
null
null
p_efbdfd6323
[]
null
20261024T182827Z_robu4a
How does LoRA differ from full fine-tuning?
How does LoRA differ from full fine-tuning?
3.158
[ "tech", "ai_interactions", "target_word" ]
[ "LoRA" ]
pw_1qt0w2y
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
LoRA
How does LoRA differ from full fine-tuning?
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261024T134105Z_no661v
null
null
null
null
null
null
null
null
0.9
-4.4
-53.3
-17.5
35.7
0
-40.5
-41.6
false
2026-10-24T18:28:27.199Z
null
null
p_69ba95c8dd
[]
null
20261024T182849Z_7gcdoa
Switch the fine-tuning run over to Unsloth before you touch the dataset again.
Switch the fine-tuning run over to Unsloth before you touch the dataset again.
4.721
[ "tech", "ai_interactions", "target_word" ]
[ "Unsloth" ]
pw_1tuuimd
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
Unsloth
Switch the fine-tuning run over to Unsloth before you touch the dataset again.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261024T134105Z_no661v
null
null
null
null
null
null
null
null
0.98
-4.2
-53.1
-18.1
35.1
0
-39.9
-41.7
false
2026-10-24T18:28:49.540Z
null
null
p_8f8676664c
[]
null
20261024T182959Z_mlfyoq
Pull the latest Ollama model before you start refactoring that service.
Pull the latest Ollama model before you start refactoring that service.
4.027
[ "tech", "ai_interactions", "target_word" ]
[ "Ollama" ]
pw_9qqkcb
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
Ollama
Pull the latest Ollama model before you start refactoring that service.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261024T134105Z_no661v
null
null
null
null
null
null
null
null
1.14
-3.7
-52
-18
34
0
-40.1
-42
false
2026-10-24T18:29:59.347Z
null
null
p_c619ffe3cf
[]
null
20261025T103753Z_6xdy4v
Stop being agentic and just fix the failing test.
Stop being agentic and just fix the failing test.
2.633
[ "tech", "ai_interactions", "target_word" ]
[ "agentic" ]
pw_157n3gi
on-demand
manual
bank
false
ThinkPad E15 Gen 3
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
agentic
Stop being agentic and just fix the failing test.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T103542Z_2ir7um
null
null
null
null
null
null
null
null
2.54
-19.8
-76.1
-35.5
40.7
0
-26
-67.4
false
2026-10-25T10:37:53.472Z
null
null
p_79b10d9b12
[]
null
20261025T103907Z_xhoxb0
Check the git log for the commit that introduced abliteration to this repo.
Check the git log for the commit that introduced abliteration to this repo.
4.198
[ "tech", "ai_interactions", "target_word" ]
[ "abliteration" ]
pw_r1grw4
on-demand
manual
bank
false
ThinkPad E15 Gen 3
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
abliteration
Check the git log for the commit that introduced abliteration to this repo.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T103542Z_2ir7um
null
null
null
null
null
null
null
null
2.18
-19.4
-75.8
-36.3
39.6
0
-23.7
-65.2
false
2026-10-25T10:39:07.089Z
null
null
p_dbcbfdc40d
[]
null
20261025T104358Z_gcdkug
I want a few-shot setup here, not zero-shot, so add three samples.
I want a few-shot setup here, not zero-shot, so add three samples.
4.334
[ "tech", "ai_interactions", "target_word" ]
[ "few-shot" ]
pw_1hc517z
on-demand
manual
bank
false
ThinkPad E15 Gen 3
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
few-shot
I want a few-shot setup here, not zero-shot, so add three samples.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T103542Z_2ir7um
null
null
null
null
null
null
null
null
0.7
-16.6
-74.9
-32.5
42.4
0
-50.8
-67.4
false
2026-10-25T10:43:58.197Z
null
null
p_b81ce25ac7
[]
null
20261025T104435Z_3f6y14
I think Muse Spark got confused by the two config files, so check that.
I think Muse Spark got confused by the two config files, so check that.
4.054
[ "tech", "ai_interactions", "target_word" ]
[ "Muse Spark" ]
pw_19dgeoe
on-demand
manual
bank
false
ThinkPad E15 Gen 3
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
Muse Spark
I think Muse Spark got confused by the two config files, so check that.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T103542Z_2ir7um
null
null
null
null
null
null
null
null
1.46
-18.1
-74.5
-35
39.5
0
-30
-62.3
false
2026-10-25T10:44:35.913Z
null
null
p_1b34a78347
[]
null
20261025T105045Z_jlpte8
Can you swap our local diffusion endpoint for fal and rerun the test?
Can you swap our local diffusion endpoint for fal and rerun the test?
4.838
[ "tech", "ai_interactions", "target_word" ]
[ "fal" ]
pw_twvo4t
on-demand
manual
bank
false
ThinkPad E15 Gen 3
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
null
Can you swap our local diffusion endpoint for fal and rerun the test?
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T103542Z_2ir7um
null
null
null
null
null
null
null
null
0.74
-14.5
-75.6
-33.4
42.2
0
-57.8
-67.4
false
2026-10-25T10:50:45.026Z
null
null
p_f3286e94e7
[]
null
20261025T105335Z_3bo97m
Before merging that pull request, show me the evals for the new prompt.
Before merging that pull request, show me the evals for the new prompt.
2.591
[ "tech", "ai_interactions", "target_word" ]
[ "evals" ]
pw_esx8yb
on-demand
manual
bank
false
ThinkPad E15 Gen 3
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
evals
Before merging that pull request, show me the evals for the new prompt.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T103542Z_2ir7um
null
null
null
null
null
null
null
null
2.4
-17.3
-76.7
-33.2
43.4
0
-27.2
-67.4
false
2026-10-25T10:53:35.575Z
null
null
p_78d75d0dce
[]
null
20261025T110201Z_8yxosg
Use LoRA to cut memory usage on this model.
Use LoRA to cut memory usage on this model.
3.595
[ "tech", "ai_interactions", "target_word" ]
[ "LoRA" ]
pw_1pl6i1p
on-demand
manual
bank
false
ThinkPad E15 Gen 3
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
LoRA
Use LoRA to cut memory usage on this model.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T103542Z_2ir7um
null
null
null
null
null
null
null
null
0.82
-11.2
-75.5
-30.7
44.8
0
-52.9
-66.2
false
2026-10-25T11:02:01.599Z
null
null
p_b6932218e2
[]
null
20261025T110240Z_elxzc5
Why did you pick Opus for this task when Sonnet is cheaper and fast enough?
Why did you pick Opus for this task when Sonnet is cheaper and fast enough?
3.642
[ "tech", "ai_interactions", "target_word" ]
[ "Sonnet" ]
pw_utq7ya
on-demand
manual
bank
false
ThinkPad E15 Gen 3
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
Sonnet
Why did you pick Opus for this task when Sonnet is cheaper and fast enough?
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T103542Z_2ir7um
null
null
null
null
null
null
null
null
2.16
-16.9
-74.4
-32.3
42.2
0
-21.2
-67.4
false
2026-10-25T11:02:40.489Z
null
null
p_971187411a
[]
null
20261025T110520Z_jv8x4f
Add a TTS step to the pipeline so the assistant can read summaries aloud.
Add a TTS step to the pipeline so the assistant can read summaries aloud.
5.183
[ "tech", "ai_interactions", "target_word" ]
[ "TTS" ]
pw_1j61aes
on-demand
manual
bank
false
ThinkPad E15 Gen 3
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
TTS
Add a TTS step to the pipeline so the assistant can read summaries aloud.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T103542Z_2ir7um
null
null
null
null
null
null
null
null
0
-12.3
-76.2
-32.3
43.9
0
-69.5
-68
false
2026-10-25T11:05:20.224Z
null
null
p_fbe798a363
[]
null
20261025T110633Z_7ncfw8
Check the LLM output before you commit anything to the branch.
Check the LLM output before you commit anything to the branch.
3.826
[ "tech", "ai_interactions", "target_word" ]
[ "LLM" ]
pw_1wn9flv
on-demand
manual
bank
false
ThinkPad E15 Gen 3
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
LLM
Check the LLM output before you commit anything to the branch.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T103542Z_2ir7um
null
null
null
null
null
null
null
null
0.56
-13.6
-77
-29.9
47.1
0
-26.4
-69.5
false
2026-10-25T11:06:33.922Z
null
null
p_a93667a23d
[]
null
20261025T111518Z_g0i7hc
Why did you start writing files when we were still in plan mode?
Why did you start writing files when we were still in plan mode?
2.435
[ "tech", "ai_interactions" ]
[ "plan mode" ]
p_4919bf6b9c
generated:claude (Claude Code subagent)
generated
bank
false
ThinkPad E15 Gen 3
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
Why did you start writing files when we were still in plan mode?
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T103542Z_2ir7um
null
null
null
null
null
null
null
null
1.82
-18.5
-74.1
-32.4
41.7
0
-18.5
-64.7
false
2026-10-25T11:15:18.952Z
null
null
p_4919bf6b9c
[]
null
20261025T111530Z_z0s8ei
Please run DeepSeek on that failing test and show me the error log.
Please run DeepSeek on that failing test and show me the error log.
4.255
[ "tech", "ai_interactions", "target_word" ]
[ "DeepSeek" ]
pw_6uhbzd
on-demand
manual
bank
false
ThinkPad E15 Gen 3
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
null
Please run DeepSeek on that failing test and show me the error log.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T103542Z_2ir7um
null
null
null
null
null
null
null
null
0
-19.3
-77.3
-34.9
42.4
0
-63.5
-70.3
false
2026-10-25T11:15:30.164Z
null
null
p_b1da3d9ea2
[]
null
20261025T111718Z_vguh3c
Is Cline set up with the same MCP servers as Claude Code?
Is Cline set up with the same MCP servers as Claude Code?
3.829
[ "tech", "ai_interactions" ]
[ "Cline" ]
p_443707214e
generated:claude (Claude Code subagent)
generated
bank
false
ThinkPad E15 Gen 3
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
Is Cline set up with the same MCP servers as Claude Code?
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T103542Z_2ir7um
null
null
null
null
null
null
null
null
0.64
-15
-75.5
-31.7
43.8
0
-68
-68
false
2026-10-25T11:17:18.255Z
null
null
p_443707214e
[]
null
20261025T111922Z_urzzne
Why did you skip the tokenizer download from Hugging Face?
Why did you skip the tokenizer download from Hugging Face?
3.705
[ "tech", "ai_interactions", "target_word" ]
[ "Hugging Face" ]
pw_vqg0lh
on-demand
manual
bank
false
ThinkPad E15 Gen 3
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
null
Why did you skip the tokenizer download from Hugging Face?
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T103542Z_2ir7um
null
null
null
null
null
null
null
null
0.54
-11.8
-76.6
-32.4
44.3
0
-56
-68
false
2026-10-25T11:19:22.114Z
null
null
p_f8d1490651
[]
null
20261025T112024Z_8cekuz
Give it a bigger thinking budget, it keeps rushing the plan.
Give it a bigger thinking budget, it keeps rushing the plan.
4.735
[ "tech", "ai_interactions" ]
[ "thinking budget" ]
p_b16b8f9d6c
generated:claude (Claude Code subagent)
generated
bank
false
ThinkPad E15 Gen 3
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
Give it a bigger thinking budget, it keeps rushing the plan.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T103542Z_2ir7um
null
null
null
null
null
null
null
null
0
-14.3
-77.1
-33.6
43.5
0
-67.4
-66.2
false
2026-10-25T11:20:24.941Z
null
null
p_b16b8f9d6c
[]
null
20261025T120141Z_88l5kd
Is there a skill for this already, or are you improvising the workflow?
Is there a skill for this already, or are you improvising the workflow?
3.783
[ "tech", "ai_interactions" ]
[ "skill" ]
p_256c798a77
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
skill
Is there a skill for this already, or are you improvising the workflow?
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T120113Z_kd1tsz
null
null
null
null
null
null
null
null
0.84
-2.7
-39.2
-15.8
23.4
0
-26.9
-32
false
2026-10-25T12:01:41.016Z
null
null
p_256c798a77
[]
null
20261025T120201Z_empa6r
The skill told you to fetch master first, so why didn't you?
The skill told you to fetch master first, so why didn't you?
2.473
[ "tech", "ai_interactions" ]
[ "skill" ]
p_d864ea8635
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
The skill told you to fetch master first, so why didn't you?
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T120113Z_kd1tsz
null
null
null
null
null
null
null
null
1.58
-5.9
-37
-16.7
20.3
0
-29.3
-23.7
false
2026-10-25T12:02:01.975Z
null
null
p_d864ea8635
[]
null
20261025T120308Z_jo3gt9
Use Claude in Chrome to check the page on a phone-sized window.
Use Claude in Chrome to check the page on a phone-sized window.
3.547
[ "tech", "ai_interactions" ]
[ "Claude in Chrome" ]
p_410d00a975
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
Claude in Chrome
Use Claude in Chrome to check the page on a phone-sized window.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T120113Z_kd1tsz
null
null
null
null
null
null
null
null
0.66
-2.6
-39.4
-14.8
24.6
0
-28.1
-30.1
false
2026-10-25T12:03:08.906Z
null
null
p_410d00a975
[]
null
20261025T122535Z_6v74un
Why did you go all agentic on a one-line change?
Why did you go all agentic on a one-line change?
3.045
[ "tech", "ai_interactions", "target_word" ]
[ "agentic" ]
pw_w51qsv
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
agentic
Why did you go all agentic on a one-line change?
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T120113Z_kd1tsz
null
null
null
null
null
null
null
null
0.84
-3.2
-40.6
-14.1
26.5
0
-32.2
-30.9
false
2026-10-25T12:25:35.714Z
null
null
p_17e87c0f4b
[]
null
20261025T123628Z_b0fxdo
We need to test the abliteration pipeline on a smaller model before scaling up.
We need to test the abliteration pipeline on a smaller model before scaling up.
4.731
[ "tech", "ai_interactions", "target_word" ]
[ "abliteration" ]
pw_113c6f
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
abliteration
We need to test the abliteration pipeline on a smaller model before scaling up.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T120113Z_kd1tsz
null
null
null
null
null
null
null
null
1.06
-7.3
-50.4
-19.9
30.5
0
-40.6
-43.3
false
2026-10-25T12:36:28.469Z
null
null
p_cc9a40cc1d
[]
null
20261025T130328Z_iim1bc
Go ahead and Muse Spark the release notes once the tests pass.
Go ahead and Muse Spark the release notes once the tests pass.
3.731
[ "tech", "ai_interactions", "target_word" ]
[ "Muse Spark" ]
pw_10xlnlx
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
Muse Spark
Go ahead and Muse Spark the release notes once the tests pass.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T120113Z_kd1tsz
null
null
null
null
null
null
null
null
1.02
-5.2
-41.6
-15.3
26.3
0
-33.5
-30.7
false
2026-10-25T13:03:28.559Z
null
null
p_0aefe95940
[]
null
20261025T140638Z_lxl307
Add a fal API key to the environment before you start the batch job.
Add a fal API key to the environment before you start the batch job.
4.149
[ "tech", "ai_interactions", "target_word" ]
[ "fal" ]
pw_1pjwx3g
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
null
Add a fal API key to the environment before you start the batch job.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T120113Z_kd1tsz
null
null
null
null
null
null
null
null
0.96
-5.4
-40.9
-17.3
23.6
0
-30.6
-31.9
false
2026-10-25T14:06:38.889Z
null
null
p_73b77a47be
[]
null
20261025T141205Z_bx3nsq
Our evals keep failing on the long context cases, so dig into those first.
Our evals keep failing on the long context cases, so dig into those first.
4.777
[ "tech", "ai_interactions", "target_word" ]
[ "evals" ]
pw_1d7awkq
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
evals
Our evals keep failing on the long context cases, so dig into those first.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T120113Z_kd1tsz
null
null
null
null
null
null
null
null
0.94
-6.5
-43
-20.2
22.8
0
-34.6
-34.7
false
2026-10-25T14:12:05.024Z
null
null
p_bb890643d5
[]
null
20261025T142538Z_hhbf5j
Turn what we just did into a skill so the next session can repeat it.
Turn what we just did into a skill so the next session can repeat it.
3.601
[ "tech", "ai_interactions" ]
[ "skill" ]
p_c6ec6c3476
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
skill
Turn what we just did into a skill so the next session can repeat it.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T141545Z_6wrz03
null
null
null
null
null
null
null
null
0.82
-10.8
-42.5
-20.9
21.6
0
-35.5
-30.8
false
2026-10-25T14:25:38.849Z
null
null
p_c6ec6c3476
[]
null
20261025T142605Z_gttbzn
After the abliteration step, the model should respond without the usual refusals.
After the abliteration step, the model should respond without the usual refusals.
4.025
[ "tech", "ai_interactions", "target_word" ]
[ "abliteration" ]
pw_otxkrg
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
abliteration
After the abliteration step, the model should respond without the usual refusals.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T141545Z_6wrz03
null
null
null
null
null
null
null
null
1.5
-9.4
-42.8
-22.2
20.5
0
-23.5
-34.5
false
2026-10-25T14:26:05.284Z
null
null
p_1ca2fde632
[]
null
20261025T142814Z_oi1ttd
Stop guessing and switch to a few-shot approach with my earlier commits.
Stop guessing and switch to a few-shot approach with my earlier commits.
3.125
[ "tech", "ai_interactions", "target_word" ]
[ "few-shot" ]
pw_13a7tf2
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
few-shot
Stop guessing and switch to a few-shot approach with my earlier commits.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T141545Z_6wrz03
null
null
null
null
null
null
null
null
1.44
-8.5
-38.9
-21.8
17.1
0
-18.8
-23.6
false
2026-10-25T14:28:14.931Z
null
null
p_1a6df7523c
[]
null
20261025T142846Z_rj7o71
I want every generated asset in this pipeline coming from fal, not a stub.
I want every generated asset in this pipeline coming from fal, not a stub.
4.423
[ "tech", "ai_interactions", "target_word" ]
[ "fal" ]
pw_1hr6rx5
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
null
I want every generated asset in this pipeline coming from fal, not a stub.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T141545Z_6wrz03
null
null
null
null
null
null
null
null
1
-10
-43.7
-22.3
21.4
0
-33.4
-34.8
false
2026-10-25T14:28:46.017Z
null
null
p_c69800ee79
[]
null
20261025T143136Z_5dopq1
I want to apply LoRA to the attention layers only.
I want to apply LoRA to the attention layers only.
3.247
[ "tech", "ai_interactions", "target_word" ]
[ "LoRA" ]
pw_tt0bc3
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
LoRA
I want to apply LoRA to the attention layers only.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T141545Z_6wrz03
null
null
null
null
null
null
null
null
0.96
-6.8
-45
-19
26
0
-36.4
-34.4
false
2026-10-25T14:31:36.895Z
null
null
p_615e7ec908
[]
null
20261025T143143Z_pyukdd
Why did Claude Code rewrite the whole file instead of patching one function?
Why did Claude Code rewrite the whole file instead of patching one function?
4.153
[ "tech", "ai_interactions", "target_word" ]
[ "Claude Code" ]
pw_1vz8sux
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
Claude Code
Why did Claude Code rewrite the whole file instead of patching one function?
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T141545Z_6wrz03
null
null
null
null
null
null
null
null
0.94
-7
-43.5
-20.3
23.2
0
-32.3
-33.7
false
2026-10-25T14:31:43.048Z
null
null
p_e9dc26b3d1
[]
null
20261025T143326Z_qdqttz
Hold off on the commit until the evals finish cleanly.
Hold off on the commit until the evals finish cleanly.
3.433
[ "tech", "ai_interactions", "target_word" ]
[ "evals" ]
pw_1y1avrf
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
evals
Hold off on the commit until the evals finish cleanly.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T141545Z_6wrz03
null
null
null
null
null
null
null
null
1.02
-5.5
-44
-20.7
23.2
0
-33.9
-36.6
false
2026-10-25T14:33:26.297Z
null
null
p_afaa14c44e
[]
null
20261025T143354Z_ic9mk0
Install Unsloth first, then rerun the training script from the clean checkout.
Install Unsloth first, then rerun the training script from the clean checkout.
4.279
[ "tech", "ai_interactions", "target_word" ]
[ "Unsloth" ]
pw_1dbcpns
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
Unsloth
Install Unsloth first, then rerun the training script from the clean checkout.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T141545Z_6wrz03
null
null
null
null
null
null
null
null
1.08
-7.9
-43.7
-21.2
22.5
0
-35.1
-35.3
false
2026-10-25T14:33:54.979Z
null
null
p_9909ffe67f
[]
null
20261025T143647Z_9b101q
Open your own tab in Claude in Chrome, don't take over mine.
Open your own tab in Claude in Chrome, don't take over mine.
3.907
[ "tech", "ai_interactions" ]
[ "Claude in Chrome" ]
p_957f9e32a3
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
Claude in Chrome
Open your own tab in Claude in Chrome, don't take over mine.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T143634Z_xrupbc
null
null
null
null
null
null
null
null
1.74
-9.3
-43.1
-20.1
23
0
-32.6
-32.4
false
2026-10-25T14:36:47.520Z
null
null
p_957f9e32a3
[]
null
20261025T143755Z_whnon4
Make the repo private and put it in the audio dictation folder.
Make the repo private and put it in the audio dictation folder.
4.059
[ "tech", "ai_interactions" ]
[ "repo" ]
p_91099d2cd8
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
repo
Make the repo private and put it in the audio dictation folder.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T143634Z_xrupbc
null
null
null
null
null
null
null
null
0.74
-5.9
-42.8
-19.2
23.6
0
-33.8
-32.6
false
2026-10-25T14:37:55.661Z
null
null
p_91099d2cd8
[]
null
20261025T143930Z_1lzchb
Claude in Chrome should open that page and summarise the main points for me.
Claude in Chrome should open that page and summarise the main points for me.
3.558
[ "tech", "ai_interactions", "target_word" ]
[ "Claude in Chrome" ]
pw_5tu1g
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
Claude in Chrome
Claude in Chrome should open that page and summarise the main points for me.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T143923Z_pi9dh1
null
null
null
null
null
null
null
null
0.5
-6.2
-44.3
-21.3
23
0
-37.7
-35.3
false
2026-10-25T14:39:30.014Z
null
null
p_e41097fbd5
[]
null
20261025T144132Z_i2a4fx
It takes real skill to read that log without losing your mind.
It takes real skill to read that log without losing your mind.
3.183
[ "tech", "ai_interactions" ]
[ "skill" ]
p_3123985fce
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
skill
It takes real skill to read that log without losing your mind.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T144023Z_cucxgt
null
null
null
null
null
null
null
null
3.24
-7.4
-45.7
-19.8
25.9
0
-36.3
-32.7
false
2026-10-25T14:41:32.425Z
null
null
p_3123985fce
[]
null
20261025T185212Z_uscv69
Can you check whether Claude in Chrome is installed on this machine yet?
Can you check whether Claude in Chrome is installed on this machine yet?
3.835
[ "tech", "ai_interactions", "target_word" ]
[ "Claude in Chrome" ]
pw_14t1wgv
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
Claude in Chrome
Can you check whether Claude in Chrome is installed on this machine yet?
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T185147Z_v85cin
null
null
null
null
null
null
null
null
1.22
-12.8
-46.2
-25.5
20.7
0
-33.6
-36.6
false
2026-10-25T18:52:12.209Z
null
null
p_8da162d00f
[]
null
20261025T185223Z_b1oger
Which repo does this script actually belong in?
Which repo does this script actually belong in?
2.762
[ "tech", "ai_interactions" ]
[ "repo" ]
p_8ba56d606e
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
repo
Which repo does this script actually belong in?
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T185147Z_v85cin
null
null
null
null
null
null
null
null
1.2
-11
-45.4
-23.4
22.1
0
-11
-34.9
false
2026-10-25T18:52:23.999Z
null
null
p_8ba56d606e
[]
null
20261025T185357Z_p2rbmg
Push a branch with your abliteration changes and open a pull request for review.
Push a branch with your abliteration changes and open a pull request for review.
4.519
[ "tech", "ai_interactions", "target_word" ]
[ "abliteration" ]
pw_4nlxal
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
abliteration
Push a branch with your abliteration changes and open a pull request for review.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T185147Z_v85cin
null
null
null
null
null
null
null
null
1.4
-9.7
-46
-21.2
24.8
0
-37.1
-36.2
false
2026-10-25T18:53:57.066Z
null
null
p_41b4196325
[]
null
20261025T185634Z_bsfdrh
Add a few-shot example for the failing test case first.
Add a few-shot example for the failing test case first.
3.495
[ "tech", "ai_interactions", "target_word" ]
[ "few-shot" ]
pw_r8sxaa
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
few-shot
Add a few-shot example for the failing test case first.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T185147Z_v85cin
null
null
null
null
null
null
null
null
0.84
-8
-46.3
-22.1
24.2
0
-35.7
-35.1
false
2026-10-25T18:56:34.007Z
null
null
p_65888e4967
[]
null
20261025T185639Z_aribzi
Do not let Muse Spark push straight to main without review.
Do not let Muse Spark push straight to main without review.
3.534
[ "tech", "ai_interactions", "target_word" ]
[ "Muse Spark" ]
pw_8qkef5
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
Muse Spark
Do not let Muse Spark push straight to main without review.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T185147Z_v85cin
null
null
null
null
null
null
null
null
0.94
-9
-45.4
-23
22.4
0
-37.1
-36.2
false
2026-10-25T18:56:39.546Z
null
null
p_0aa9a96abf
[]
null
20261025T185647Z_vdxbiu
Push the commit before you get agentic with the refactor.
Push the commit before you get agentic with the refactor.
5.79
[ "tech", "ai_interactions", "target_word" ]
[ "agentic" ]
pw_1lzeemv
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
agentic
Push the commit before you get agentic with the refactor.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T185147Z_v85cin
null
null
null
null
null
null
null
null
0.7
-9.9
-46.9
-25.8
21.1
0
-36.8
-34.7
false
2026-10-25T18:56:47.152Z
null
null
p_00af8ba41e
[]
null
20261025T185710Z_jmfflk
Check whether fal supports streaming responses for the video model.
Check whether fal supports streaming responses for the video model.
3.585
[ "tech", "ai_interactions", "target_word" ]
[ "fal" ]
pw_1heijgx
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
null
Check whether fal supports streaming responses for the video model.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T185147Z_v85cin
null
null
null
null
null
null
null
null
1.1
-7.3
-44.6
-22.1
22.5
0
-34.3
-36.5
false
2026-10-25T18:57:10.947Z
null
null
p_d872a2fe35
[]
null
20261025T185735Z_cn4tlb
Load the skill before you touch any config file.
Load the skill before you touch any config file.
3.174
[ "tech", "ai_interactions", "target_word" ]
[ "skill" ]
pw_ec93ty
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
skill
Load the skill before you touch any config file.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T185147Z_v85cin
null
null
null
null
null
null
null
null
0.74
-10.2
-45.7
-21.8
23.9
0
-34.3
-35.1
false
2026-10-25T18:57:35.591Z
null
null
p_bb5233fc05
[]
null
20261025T185824Z_95mgwd
Can you check whether the evals actually cover tool calling at all?
Can you check whether the evals actually cover tool calling at all?
4.111
[ "tech", "ai_interactions", "target_word" ]
[ "evals" ]
pw_ya5nxw
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
evals
Can you check whether the evals actually cover tool calling at all?
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T185147Z_v85cin
null
null
null
null
null
null
null
null
0.8
-8.8
-44.6
-20.4
24.3
0
-35.1
-37.3
false
2026-10-25T18:58:24.023Z
null
null
p_26cc5f07d6
[]
null
20261025T185853Z_i7mzob
Commit the config that pins Unsloth to a specific version, please.
Commit the config that pins Unsloth to a specific version, please.
3.629
[ "tech", "ai_interactions", "target_word" ]
[ "Unsloth" ]
pw_1tms2nh
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
Unsloth
Commit the config that pins Unsloth to a specific version, please.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T185147Z_v85cin
null
null
null
null
null
null
null
null
0.84
-8
-44.6
-22.9
21.6
0
-36.6
-35.1
false
2026-10-25T18:58:53.484Z
null
null
p_9cd46d7d4d
[]
null
20261025T185904Z_50ohhs
Why did the LLM rewrite that function without being asked?
Why did the LLM rewrite that function without being asked?
3.493
[ "tech", "ai_interactions", "target_word" ]
[ "LLM" ]
pw_w3w85t
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
LLM
Why did the LLM rewrite that function without being asked?
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T185147Z_v85cin
null
null
null
null
null
null
null
null
0.64
-6.4
-45.3
-22.2
23.1
0
-34.2
-35.8
false
2026-10-25T18:59:04.327Z
null
null
p_ba01cff441
[]
null
20261025T185947Z_cb155o
Make sure the ONNX opset matches the runtime version.
Make sure the ONNX opset matches the runtime version.
4.017
[ "tech", "ai_interactions", "target_word" ]
[ "ONNX" ]
pw_1ix45dx
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
ONNX
Make sure the ONNX opset matches the runtime version.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T185147Z_v85cin
null
null
null
null
null
null
null
null
0.82
-7.8
-45.5
-22
23.5
0
-37.8
-36
false
2026-10-25T18:59:47.913Z
null
null
p_d75725a697
[]
null
20261025T190113Z_5p5mps
Check whether that checkpoint is actually hosted on Hugging Face or just linked there.
Check whether that checkpoint is actually hosted on Hugging Face or just linked there.
4.809
[ "tech", "ai_interactions", "target_word" ]
[ "Hugging Face" ]
pw_fqmtgk
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
null
Check whether that checkpoint is actually hosted on Hugging Face or just linked there.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T185147Z_v85cin
null
null
null
null
null
null
null
null
0.86
-7.6
-45.4
-20.9
24.5
0
-34.7
-37.2
false
2026-10-25T19:01:13.065Z
null
null
p_81b7070880
[]
null
20261025T190132Z_qe4jcm
That transcript looks wrong, so check whether Nemotron is actually loaded.
That transcript looks wrong, so check whether Nemotron is actually loaded.
4.155
[ "tech", "ai_interactions", "target_word" ]
[ "Nemotron" ]
pw_dh7sky
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
Nemotron
That transcript looks wrong, so check whether Nemotron is actually loaded.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T185147Z_v85cin
null
null
null
null
null
null
null
null
1.22
-5.7
-45
-20.4
24.7
0
-35.1
-35.1
false
2026-10-25T19:01:32.683Z
null
null
p_3e6334d3ec
[]
null
20261025T190142Z_8rjolv
Switch the local inference engine to Qwen for this batch of tests.
Switch the local inference engine to Qwen for this batch of tests.
3.521
[ "tech", "ai_interactions", "target_word" ]
[ "Qwen" ]
pw_m00wp7
on-demand
manual
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
null
Qwen
Switch the local inference engine to Qwen for this batch of tests.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T185147Z_v85cin
null
null
null
null
null
null
null
null
1.86
-10
-44.7
-23.2
21.5
0
-15.9
-33.4
false
2026-10-25T19:01:42.643Z
null
null
p_07bbe44c99
[]
null
20261025T210118Z_743ks7
Compare what Kilo Code did with the Cline version of the same fix.
Compare what Kilo Code did with the Cline version of the same fix.
3.751
[ "tech", "ai_interactions" ]
[ "Kilo Code" ]
p_4bdc31605f
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
Compare what Kilo Code did with the Cline version of the same fix.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T205816Z_9cw3sr
null
null
null
null
null
null
null
null
0.68
-6.3
-37.6
-20.3
17.3
0
-25.2
-28.2
false
2026-10-25T21:01:18.540Z
null
null
p_4bdc31605f
[]
null
20261025T211741Z_4o2nf2
Write a PreToolUse hook that stops any force push to master.
Write a PreToolUse hook that stops any force push to master.
4.498
[ "tech", "ai_interactions" ]
[ "PreToolUse hook" ]
p_40022f5e61
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
Write a PreToolUse hook that stops any force push to master.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T205816Z_9cw3sr
null
null
null
null
null
null
null
null
0.84
-8
-40.8
-21.8
19
0
-26.1
-31.4
false
2026-10-25T21:17:41.195Z
null
null
p_40022f5e61
[]
null
20261025T211747Z_r59pvw
Don't use YOLO mode on the server, I want to see every command.
Don't use YOLO mode on the server, I want to see every command.
4.307
[ "tech", "ai_interactions" ]
[ "YOLO mode" ]
p_7758089536
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
Don't use YOLO mode on the server, I want to see every command.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T205816Z_9cw3sr
null
null
null
null
null
null
null
null
0.7
-7.3
-38.8
-21.9
16.9
0
-30.2
-27.8
false
2026-10-25T21:17:47.216Z
null
null
p_7758089536
[]
null
20261025T211754Z_l4q9qm
Run Codex in YOLO mode inside the container, it can't do any damage there.
Run Codex in YOLO mode inside the container, it can't do any damage there.
5.327
[ "tech", "ai_interactions" ]
[ "YOLO mode" ]
p_2ef93bba58
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
Run Codex in YOLO mode inside the container, it can't do any damage there.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T205816Z_9cw3sr
null
null
null
null
null
null
null
null
0.8
-7.8
-40.7
-22.3
18.4
0
-30.3
-27.8
false
2026-10-25T21:17:54.408Z
null
null
p_2ef93bba58
[]
null
20261025T212222Z_4rnlkj
Close the GitHub issue once the fix is released.
Close the GitHub issue once the fix is released.
3.501
[ "tech", "ai_interactions" ]
[ "GitHub issue" ]
p_7390902c6f
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
Close the GitHub issue once the fix is released.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T205816Z_9cw3sr
null
null
null
null
null
null
null
null
0.76
-11.2
-41.7
-23.8
18
0
-32
-27
false
2026-10-25T21:22:22.999Z
null
null
p_7390902c6f
[]
null
20261025T212741Z_ww99al
GitHub Actions is failing on the lint step, read the log.
GitHub Actions is failing on the lint step, read the log.
3.542
[ "tech", "ai_interactions" ]
[ "GitHub Actions" ]
p_c0c0b9c1de
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
GitHub Actions is failing on the lint step, read the log.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T205816Z_9cw3sr
null
null
null
null
null
null
null
null
1.38
-4.5
-54
-19.3
34.6
0
-40.8
-41.5
false
2026-10-25T21:27:41.681Z
null
null
p_c0c0b9c1de
[]
null
20261025T213352Z_yvoznj
Which feature branch has the speaker picker work on it?
Which feature branch has the speaker picker work on it?
3.465
[ "tech", "ai_interactions" ]
[ "feature branch" ]
p_3e12efae23
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
Which feature branch has the speaker picker work on it?
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T205816Z_9cw3sr
null
null
null
null
null
null
null
null
0.94
-5.2
-59.5
-21.2
38.3
0
-49.1
-47.8
false
2026-10-25T21:33:52.981Z
null
null
p_3e12efae23
[]
null
20261025T213358Z_xxwusy
Put this on a feature branch, not straight on master.
Put this on a feature branch, not straight on master.
3.203
[ "tech", "ai_interactions" ]
[ "feature branch" ]
p_e4e63af735
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
Put this on a feature branch, not straight on master.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261025T205816Z_9cw3sr
null
null
null
null
null
null
null
null
1.02
-7.6
-60
-21.3
38.7
0
-50.3
-46.3
false
2026-10-25T21:33:58.176Z
null
null
p_e4e63af735
[]
null
20261026T093046Z_17utd2
Does the thinking budget count against my usage limit?
Does the thinking budget count against my usage limit?
2.07
[ "tech", "ai_interactions" ]
[ "thinking budget" ]
p_4e6d4cceb3
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
Does the thinking budget count against my usage limit?
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261026T092249Z_2fxg7r
null
null
null
null
null
null
null
null
1.86
-8.6
-39.2
-22.2
17
0
-8.6
-29
false
2026-10-26T09:30:46.059Z
null
null
p_4e6d4cceb3
[]
null
20261026T093107Z_votmv3
Switch Roo Code to the orchestrator mode and let it plan the migration first.
Switch Roo Code to the orchestrator mode and let it plan the migration first.
4.674
[ "tech", "ai_interactions" ]
[ "orchestrator" ]
p_47764ef793
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
Switch Roo Code to the orchestrator mode and let it plan the migration first.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261026T092249Z_2fxg7r
null
null
null
null
null
null
null
null
1.08
-6.5
-38.6
-21.2
17.5
0
-8.3
-27.8
false
2026-10-26T09:31:07.843Z
null
null
p_47764ef793
[]
null
20261026T093114Z_a56ar3
You're the orchestrator on this one, so break it into pieces and hand them out.
You're the orchestrator on this one, so break it into pieces and hand them out.
4.685
[ "tech", "ai_interactions" ]
[ "orchestrator" ]
p_0790d1f4ba
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
You're the orchestrator on this one, so break it into pieces and hand them out.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261026T092249Z_2fxg7r
null
null
null
null
null
null
null
null
0.6
-5.2
-38.9
-19.3
19.6
0
-28.2
-28.9
false
2026-10-26T09:31:14.225Z
null
null
p_0790d1f4ba
[]
null
20261026T093153Z_alda1x
Just cherry-pick the fix onto master, the rest of that branch isn't ready.
Just cherry-pick the fix onto master, the rest of that branch isn't ready.
4.291
[ "tech", "ai_interactions" ]
[ "cherry-pick" ]
p_48b7708a89
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
Just cherry-pick the fix onto master, the rest of that branch isn't ready.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261026T092249Z_2fxg7r
null
null
null
null
null
null
null
null
1.26
-8.4
-38.5
-20.7
17.8
0
-8.4
-29.3
false
2026-10-26T09:31:53.892Z
null
null
p_48b7708a89
[]
null
20261026T121214Z_34evf1
Open a GitHub issue for the bug and link it in the PR.
Open a GitHub issue for the bug and link it in the PR.
3.549
[ "tech", "ai_interactions" ]
[ "GitHub issue" ]
p_d40cb340e3
generated:claude (Claude Code subagent)
generated
bank
false
OnePlus Nord 3 5G
internal
quiet-indoor
[]
null
[]
[]
false
null
null
[]
null
tech
null
Open a GitHub issue for the bug and link it in the PR.
null
null
null
null
null
null
null
null
null
null
null
Daniel Rosehill
8992
Daniel Rosehill
male
48,000
24
flac
48,000
stt
false
null
script
null
null
null
null
null
20261026T092249Z_2fxg7r
null
null
null
null
null
null
null
null
0.68
-4.4
-50.1
-16.6
33.6
0
-40.5
-37
false
2026-10-26T12:12:14.047Z
null
null
p_d40cb340e3
[]
null
End of preview. Expand in Data Studio

AI Interactions Sentences

Type: Audio Role: Split-out Subset of: Tech-Sentences

Audio dataset, split-out. Rows copied from the master, danielrosehill/Read-Aloud-Voice-Sentences, so one task can be browsed alone. Training reads the master, not this copy. Subset of danielrosehill/Tech-Sentences: its takes are all in that dataset too. Every audio dataset as one tree, with hours: the audio index. How every dataset fits together: the dataset index.

Instructions dictated to AI coding agents (Claude Code, Codex and others) and the git and GitHub workflow around them, read aloud by one speaker: subagents, worktrees, compaction, plan mode, hooks, MCP servers, pull requests, rebases. Sentences to read: AI-Interactions-Sentence-Bank.

Website: hebrish.app: get the Hebrish keyboard for Android and the desktop apps, see the results, and find every model, dataset and demo in one place.

Two tabs. sentences lists every ground-truth sentence with a takes count, so takes = 0 shows what has not been recorded. takes has one row per recording with its audio, joined on sentence_id; a sentence can have several takes. load_dataset("danielrosehill/AI-Interactions-Sentences", "sentences") reads the text; the default config reads the takes.

Takes come from danielrosehill/Read-Aloud-Voice-Sentences, the full dataset, with the same ids and train/test split, by scripts/sync_subsets.py in the project repository. Do not edit this dataset directly. New takes append and only the sentence counts are rewritten.

Speaker: Daniel Rosehill, the dataset's author, who was born in Ireland and speaks English with an Irish accent, and who lives in Jerusalem. Every recording is his voice. Deliberate variation: he sometimes changes his speaking speed on purpose (fast and rushed, or slow) and sometimes says a word the way people commonly mispronounce it, to make the models robust to real speech. The transcript is always the sentence as written, so a take whose speed or pronunciation sounds off against its card is usually intentional, not a recording error.

All audio was recorded on a OnePlus Nord 3 5G phone using its internal microphone, with browser echo cancellation, noise suppression and auto gain turned off.

Layout

sentences/train.parquet  one row per sentence: `id`, `text`, `takes`, `prompt_ids`, `served` and the sentence columns
data/train-*.parquet   one row per take: `audio` (the FLAC or WAV file, embedded) and every column
meta/train.jsonl       index: the same rows without audio, plus `shard`, `audio_path` and the audio's byte range
meta/schema.json       column types
data/test-*.parquet, meta/test.jsonl   held-out set, never trained on

Audio is 48 kHz mono 24-bit FLAC for rows from 2026-10-22 and 16 kHz mono 16-bit WAV before.

The test split is term-disjoint: a held-out term never appears in train. load_dataset("danielrosehill/AI-Interactions-Sentences") works directly.

Columns

A take's sentence_id is its row's id in sentences. A sentence with served: false is no longer offered by the app; it is kept because a take used it. The columns are the same as in the main dataset: text (ground truth, punctuated and cased like Parakeet v2 output), text_he (the same transcript with its Hebrew terms in Hebrew script, for Hebrew-capable models), categories, target_terms, prompt_source, prompt_origin (manual / dataset / generated), edited, duration_s, environment, conditions, device, speaker, speaker_gender, contributor, hebrew_pronunciation, variant_groups, variant_spelling and recorded_at. The main dataset's card describes them.

Licence

CC-BY-4.0, with sentence-text attribution per prompt_source as described in the main dataset's licence section.

Downloads last month
1,678

Spaces using danielrosehill/AI-Interactions-Sentences 2

Collection including danielrosehill/AI-Interactions-Sentences