Harry Shum

Head of AI & Research, Microsoft · 0 appearances on the record.

computed by AI from the episodes · how this works → · full disclaimer →

14statements → 10claims → 0claims resolved → average certainty → average debate potential →

7 not checkable as stated how the 10 claims stand · each chip opens the sources

10 assertions · 3 opinions · 1 insight · every statement was checked. The predictions and assertions are the 10 claims: statements the public record can support or contradict. 3 are resolved, and 7 name no date, number or outcome precise enough to check. Everything else (opinions, insights, what ifs, disclosures) can never be settled by the record, so it carries no assessment.

The record, in short

What the tape says about how Harry argues and how the claims held up. Everything they said, and everything said about them, is in the tabs below.

Their most notable supported claim

Assertion Supported
Shum: Microsoft Research Beijing invented ResNet, the most popular vision network
“ResNet now is the most popular deep neural net, ah, in computer vision. Ah, we actually invented in our research lab in Beijing by my students, ah, using 152 layers of neural networks.”
Harry Shum Jan 16, 2020 ▶ 0:43 Explaining AI

Expressed certainty vs assessment result

none yet certainty 1
none yet certainty 2
none yet certainty 3
75% certainty 4
50% certainty 5

weighted support: a fully supported claim counts one, a partly supported claim counts half. Each filled bar is clickable and opens exactly those claims; "none yet" means nothing said at that certainty level has resolved yet

How they sound: speaking style how? →

226 words/min while actually speaking · 49.7 um and uh per 1k words

No argument clarity score for Harry Shum: no usable question→answer exchanges on raw tape (a fair score needs 8+). We do not score a sample that small. Roundtable and news formats yield far fewer direct exchanges than interviews.

Measured by listening to the audio itself: 2,433 words across 1 episode of raw-level tape, transcribed verbatim with every um and uh kept, each one attributed only where the alignment onto our timed stream is unambiguous. These are measurements of speaking style. We do not rank them: across this corpus, fluency and argument quality are nearly uncorrelated (ρ≈0.2), and smooth talking does not signal clear thinking. How it's measured →

Everything Harry Shum said on the a16z Podcast that made the record, most notable first. Filter by type, assessment or year in the ledger →

Opinion
Shum: Opaque AI in medical and military decisions can be deadly
“You might think it's probably harmless to get some kind of Netflix, Netflix recommendation about the movies you to watch. But the political ads on social networks can be problematic. Medical diagnosis, or even military decisions based on AI, you know, can be d…”
Harry Shum Jan 16, 2020 ▶ 13:32 Explaining AI
Assertion Partly supported
Shum: Microsoft AI surpassed human transcribers with 5.1% error rate
“In speech, ah, there's this very standard dataset called the switchboard dataset that is really recording of two, ah, from two sides and, ah, phone conversations. Ah, two years ago, we accomplished the 5.1% of the error rate. Ah, that is really, really, ah, am…”
Harry Shum Jan 16, 2020 ▶ 1:21 Explaining AI
Assertion Not checkable as stated
Shum: XiaoIce distributed 1M coupons with 40% in-store conversion rate
“So we did experiment with the coupons, and the amazing that, you know, we distributed a million coupons in 13 hours, and the in-stock conversion rate hit a 40% in the next four days.”
Harry Shum Jan 16, 2020 ▶ 5:38 Explaining AI
Assertion Partly supported
Shum: XiaoIce powers 90% of Chinese corporate earnings summaries
“We actually now power 90% of, ah the quarterly earning report summary, ah, for 90% of, you know, Chinese companies, and they're also powered by Xiaoice.”
Harry Shum Jan 16, 2020 ▶ 6:01 Explaining AI
Opinion
Shum: Goldman Sachs's Apple Card system likely contains AI bias
“You probably would, you know, agree with me that very likely there's some kind of AI buyers in the machine learning system they have built.”
Harry Shum Jan 16, 2020 ▶ 6:55 Explaining AI
Opinion
Shum: Society cannot accept unexplainable AI decision-making
“And because I really believe we cannot accept a future where AI making decisions, ah, that we cannot explain, we cannot understand.”
Harry Shum Jan 16, 2020 ▶ 16:02 Explaining AI
Assertion Not checkable as stated
Shum: Microsoft AI approaches human parity in perception and language
“We actually gradually approaching human parody in a number of those human tasks, especially perception. From computer vision to speech to more and more natural language.”
Harry Shum Jan 16, 2020 ▶ 0:26 Explaining AI
Insight
Shum: Blindly training AI on internet data yields biased systems
“The data we collected over the internet has some inherent bias. So if we just blindly use the data to train AI, we know we have problems.”
Harry Shum Jan 16, 2020 ▶ 11:15 Explaining AI
Assertion Not checkable as stated
Shum: AI models with trillions of parameters are black boxes
“Because we now train all this kind of complex models with millions, even trillions of parameters. So effectively we are building AIs as black boxes.”
Harry Shum Jan 16, 2020 ▶ 12:31 Explaining AI
Assertion Not checkable as stated
Shum: We are the first generation of humans living with AI
“Ah, we are going to be the first generation of humans to ever live with AI.”
Harry Shum Jan 16, 2020 ▶ 15:43 Explaining AI
Assertion Supported
Shum: Microsoft Research Beijing invented ResNet, the most popular vision network
“ResNet now is the most popular deep neural net, ah, in computer vision. Ah, we actually invented in our research lab in Beijing by my students, ah, using 152 layers of neural networks.”
Harry Shum Jan 16, 2020 ▶ 0:43 Explaining AI
Assertion Not checkable as stated
Shum: Microsoft's XiaoIce social chatbot has 120 million monthly active users
“Ah, XiaoEyes now is very popular, ah, with a hundred twenty million, ah, monthly active users.”
Harry Shum Jan 16, 2020 ▶ 2:47 Explaining AI
Assertion Not checkable as stated
Shum: Microsoft's XiaoIce chatbot averages 23 conversation turns per session
“But Xiaomi's CPS, or conversation turns possession, No, on average, actually achieves 23 times.”
Harry Shum Jan 16, 2020 ▶ 4:21 Explaining AI
Assertion Not checkable as stated
Shum: Longest XiaoIce conversation lasted over 29 hours and 7,000 turns
“Incredibly, the longest conversation any human user ever talked to XiaoAis is actually over 29 hours of 7000 turns.”
Harry Shum Jan 16, 2020 ▶ 4:32 Explaining AI

Played on the show (1)

Episodes where a recording of Harry Shum was played rather than Harry taking part, or where the tape carries an address with nobody putting questions to them. Listed because the words are on the record, kept out of every score on this page because they were not said on this show. We read this off the tape: who was spoken to, who was asked something, who answered whom.

EpisodeDateOn tapeWhat it is
Explaining AI Jan 16, 2020 13m aired address
Made with StarZero

Turn any episode into a week of clips.

This entire site, over 1,000 episodes transcribed, diarized, checked and made playable, runs on the StarZero media pipeline. Drop in your own episode and the podcast clipper finds the moments worth sharing, cuts them, captions them, and reframes them for every feed.