J. Kolb

5.9K posts

J. Kolb banner
J. Kolb

J. Kolb

@kolb

@[email protected]

Denver, CO Katılım Eylül 2013
612 Takip Edilen815 Takipçiler
J. Kolb
J. Kolb@kolb·
100%. It's worse than Tennyson, which most folks don't drive on to just pass through, unlike both Lowell and 32nd.
English
0
1
1
595
J. Kolb
J. Kolb@kolb·
@lowtheband what’s a song (or two) of yours from a recent record that you just love for how the effects, recording, and mixing come together?
English
0
0
2
19.9K
J. Kolb
J. Kolb@kolb·
@unfard shocked we haven't gotten any lightfoot hot takes from you
English
0
0
0
71
J. Kolb
J. Kolb@kolb·
@Zim303Rebecca 1. Berkeley 2. I feel very safe 3. My back windshield was bashed in, but I think that was 5-6yrs ago.
English
0
0
1
304
J. Kolb retweetledi
André David
André David@DrAndreDavid·
30 years ago it wasn't clear inner trackers would work at the LHC and muons were one of the few sure bets. @CMSExperiment’s “M” also reflects that and from day one it yielded beautiful collision results. See the η peak peeking out? 1/ @CmsMuon twiki.cern.ch/twiki/bin/view…
André David tweet media
English
2
8
47
8K
J. Kolb
J. Kolb@kolb·
Same
English
1
0
3
0
J. Kolb
J. Kolb@kolb·
@mnolangray It's a welcome change, but there remains a substantial lack of ground-level commercial presence, and of course the on-going emphasis on single-occupancy vehicles.
English
0
0
13
0
J. Kolb retweetledi
𝙽𝚒𝚌𝚔 𝙼𝚊𝚝𝚑𝚎𝚜𝚘𝚗
Our office kitchen contract won't be renewed after Friday. They spoiled us beyond belief. I am sad that tech workers get multi-month severance and those who serve us get nada. If anyone wants to join in a show of appreciation DM me. Hoping to do something by Friday.
English
1
3
50
0
J. Kolb
J. Kolb@kolb·
But if we can’t even begin to reason about a highly-simplified and approximate confusion matrix for these models, then how can we ever reliably use them for decision-making? How can we even begin to approximate the ways chatgpt gets things wrong, and the costs therein? 2/2
English
0
0
3
0
J. Kolb
J. Kolb@kolb·
Large language models aren’t technically classifiers, but we frequently treat them as such. We expect them to be useful by delivering some sort of “right answer”. 1/2
English
1
0
1
0
J. Kolb
J. Kolb@kolb·
Chashu in prod
J. Kolb tweet media
English
2
0
6
0
J. Kolb
J. Kolb@kolb·
Chashu!
J. Kolb tweet media
English
2
0
8
0