Brady Fukumoto

1.3K posts

Brady Fukumoto banner
Brady Fukumoto

Brady Fukumoto

@br80io

Fighting for fact fluency since 2009

San Diego, CA Katılım Mart 2014
555 Takip Edilen989 Takipçiler
Sabitlenmiş Tweet
Brady Fukumoto
Brady Fukumoto@br80io·
Most folks who think they are bad at math are really just bad at arithmetic. "I used to be good at math," they muse. "I got As in elementary but wasn't smart enough for fractions (or long division or algebra)" It's not your fault, it's the flawed system. Here's proof. (1/5)
English
11
41
226
0
Brady Fukumoto retweetledi
Anna Stokke
Anna Stokke@rastokke·
If students struggle with fraction arithmetic, algebra is inaccessible. If they struggle with basic number operations (including times tables), fractions are inaccessible. Math is hierarchical & gaps compound. There's no way around it. We have to get math right in primary.
English
83
390
2.2K
62.4K
Brady Fukumoto retweetledi
Kyle Aretae
Kyle Aretae@aretae·
@garrytan 10k hours tutoring + teaching math. ~All failure is not having 2 steps before automatic. Near 100% of fraction failure is because multiplication is too slow. MOST algebra failure is fraction failure. Most of math comprehension sits on multiplication tables.
English
4
13
67
2.5K
Brady Fukumoto retweetledi
Spencer A. Klavan
Spencer A. Klavan@SpencerKlavan·
It’s literally the exact opposite of this. Kids who pass off foundational cognitive tasks like memorization to AI will be lost in an ocean of people just like them, all powerless to think their own thoughts, dependent on bad mechanical imitations of mental acts they have no capacity to perform or judge for themselves. They’ll grow up into glazed-over subaltern dupes at the mercy of machinists who view them as little more than farm animals to milk for training data. You could hardly do a worse disservice to a young person right now than to empty out the contents of their soul and strip them of the mental armor that only a rigorous literary education can provide. And all in the name of some gullible claptrap about humanity and tech that wouldn’t stand up to five minutes’ scrutiny if the people peddling it and swilling it down had ever read a single thing worth reading. We had all better snap out of this kookery right the heck now or we’re cooked, fam.
Julia McCoy@JuliaEMcCoy

We are sending our kids to school to memorize facts that AI can retrieve in 0.3 seconds. We're grading them on essays that AI writes better than their teachers. We're preparing them for jobs that won't exist by the time they graduate. The entire education system is training humans to compete with machines at what machines do best. That's not education. That's sabotage. The schools that survive will teach thinking, not memorizing. Creating, not repeating. Discerning, not obeying. Every other school is a museum that doesn't know it yet.

English
185
1.1K
7.6K
314.4K
Brady Fukumoto retweetledi
Tom Bennett OBE
Tom Bennett OBE@tombennett71·
This is a contemporary permutation of the 'why do I need to teach things that kids can just Google?' A car can get you somewhere faster than walking, so why bother learning to walk? Because you need to walk to the car. Walking takes you to places cars can't reach. If you travelled everywhere by car, you would think the world was just a connected grid of roads and motorways. Your world would be diminished, and eventually people would forget there was even anything but roads. One reason we teach children art, science, literacy etc, so that they understand how to use the tools we have created. If we outsource thinking to AI then we lose the capacity to think. We need to use AI to help our thinking, not replace it. If you don't run the room, the room runs you.
Brendan McCord 🏛️ x 🤖@Brendan_McCord

LLMs are living off the moral and intellectual capital of a pre-AI world, just like Nietzsche said secular liberals live off Christianity. What happens when the inheritance runs out? Using LLMs well — knowing when to trust them, how to interrogate their outputs, what questions are worth asking — depends on capacities that are pre-LLM in origin: critical judgment, domain expertise, philosophical seriousness, taste. People who use LLMs well right now tend to be people formed by traditions of deep reading, argument, and intellectual discipline that were not themselves produced by or optimized for interaction with language models. The tool works for them because they bring something the tool cannot supply. Nietzsche thought secular liberals were coasting on the fumes of a Christian metaphysics they'd officially abandoned. The shadow of God lingering on the cave wall. The question is whether LLM-native thinking is the same kind of afterglow.

English
7
45
222
14.1K
Brady Fukumoto retweetledi
Ms. Sam
Ms. Sam@SciInTheMaking·
I am a high school chemistry teacher. For years, I encountered students who did not know their multiplication tables and it interfered with their ability to understand chemistry. This year, I decided to take matters into my own hands. ⬇️
English
153
260
4.3K
743.3K
Brady Fukumoto retweetledi
Brian Poncy
Brian Poncy@brian_poncy·
There have been multiple times people have used facts on fire, showed improved student scores, and kids were excited about math. The district leadership learns about it and despite the data tell teachers that they cannot use timed practice in the district. Its maddening that teachers with data are not being listened to.
English
4
13
59
4.8K
Brady Fukumoto
Brady Fukumoto@br80io·
@Austen Learning to ship software is not the same as learning to code.
English
0
0
0
18
Austen Allred
Austen Allred@Austen·
Every "learn to code" program ever created was, for obvious reasons, first designed to teach all of the syntax, low level operations, and work your way up by piecing them together. It might make sense for "learn to code" to now start with system design and architecture.
English
69
24
415
16.4K
Brady Fukumoto retweetledi
Carl Hendrick
Carl Hendrick@C_Hendrick·
Most education apps love passive formats because they're low-friction. Users don't get frustrated. Session times stay high. Paying parents are happy. Many prioritise engagement streaks. But session time ≠ learning time. An hour of clicking, watching, and recognising can involve almost zero actual cognitive application.
English
2
4
70
5.1K
Brady Fukumoto retweetledi
Dale Chu
Dale Chu@Dale_Chu·
Math scores didn’t collapse overnight. They collapsed because we treated memorization as a sin and cluttered the curriculum with “strategies.” Reading got its science moment; math is overdue.
Dale Chu tweet media
English
113
222
1.6K
64.2K
Brady Fukumoto retweetledi
Justin Skycak
Justin Skycak@justinskycak·
Yes, students need to memorize their math facts. Lack thereof is characteristic of mathematical disability. When facts are not memorized, recomputing them on the fly steals brainpower away from higher-level thought. It's quite literally disabling. And completely avoidable.
Justin Skycak tweet media
English
16
86
646
29.7K
Brady Fukumoto retweetledi
Kelley K
Kelley K@KelleyKga·
The fact is that high schools are graduating kids with As and Bs in advanced math courses who haven't mastered foundational skills. The data from the UCSD report makes that clear. 20% took calculus in high school! Their GPA in math classes is ~3.6! This is happening all over the country. Kids are over-reliant on calculators and generous grading policies. Because the UC system does test-blind admissions, these kids look like skilled math students. (Applicants are not allowed to even submit ACT or SAT scores at UC schools.) Their foundational math gaps are only identified when they take the freshman math placement exam (no calculator allowed) and find out they aren't qualified for pre-calc or higher math at the collegiate level.
Kelley K tweet mediaKelley K tweet media
Kelley K@KelleyKga

This is the part I think a lot of people don't understand. These kids are getting A's and B's in classes like Algebra II and Pre-Calculus in high school, but grade inflation and lack of standardized testing for admissions hide the issues.

English
101
232
1.5K
976.1K
Brady Fukumoto retweetledi
Brady Fukumoto retweetledi
Justin Skycak
Justin Skycak@justinskycak·
Yes, you need to be automatic on math facts. Not just arithmetic, but also exponent/logarithm rules, trig identities, derivative/integral rules, etc., all the way up the ladder of math. Why? Because solving problems feels smooth when you're automatic on the components – it’s like moving puzzle pieces around, and you just need to identify how they fit together. But without automaticity on the components, each puzzle piece is a heavy weight. You struggle to move them at all, much less figure out where they’re supposed to go. And this deficit compounds like a high-interest debt. To develop automaticity, it's critical to engage in proper retrieval practice, doing your best to retrieve information from memory. Always try your best to recall formulas and approaches without looking back at reference material. Use the reference material when you absolutely need to, but no more than you absolutely need to. If you can't retrieve a piece of info from memory despite trying your hardest, then okay, peek back at the reference – just a little bit, just the tiniest bit of priming, just that specific piece of info that you were trying to remember, nothing else – and then close the reference, re-pull the info from memory, and try to recall the rest and proceed forward as far as possible without peeking back at the reference again. Your brain is lifting a weight and the reference material is your spotter – it’s there as a last resort to help you get the weight up, only when you absolutely can’t get it up yourself, and the amount of help should be kept to the bare minimum. The goal is to wean yourself off of reference material, using it as sparsely as possible, until you don’t need it at all. That said, in addition to working problems with minimal reliance on reference material and frequently taking closed-book timed quizzes, there are plenty of instances where it's possible to efficiently speed up the development of automaticity *even further* by engaging in rapid-fire flashcard-style retrieval practice. Rapid-fire flashcard-style automaticity training is absolutely critical for multiplication tables and it also works great for trig identities, derivative/integral rules, etc. It's pretty high on my and @exojason's to-do list.
English
8
11
130
10.4K
Brady Fukumoto retweetledi
Justin Skycak
Justin Skycak@justinskycak·
Automaticity is necessary for creativity.
Justin Skycak tweet media
English
2
19
146
12.7K
Brady Fukumoto retweetledi
Sarah Farrell
Sarah Farrell@SarahFarrellKS2·
A good knowledge of times tables is so important for accessing the majority of the maths curriculum fluently and confidently. Here’s a poster showing some uses! What else would you add from the primary curriculum? What about the secondary curriculum? primarymathshub.com/everyday-essen…
Sarah Farrell tweet media
English
9
60
332
37.5K
Brady Fukumoto retweetledi
Justin Skycak
Justin Skycak@justinskycak·
The single most important word in the science of learning: "automaticity" - the ability to carry out basic skills effortlessly without conscious thought. Without automaticity, you have to consciously think about every low-level action, which overloads your working memory and leaves no room for higher-level reasoning.
English
9
71
497
49.5K
Brady Fukumoto retweetledi
Justin Skycak
Justin Skycak@justinskycak·
The greatest breakthrough in the science of learning over the last century:
Justin Skycak tweet media
English
29
217
1.8K
385.6K
Brady Fukumoto
Brady Fukumoto@br80io·
@stevemagness Steve, are you familiar with the Ecological Dynamics approach to training? Essentially, replacing rigid drills with constraint-based games in a live environment for skill acquisition. If so, I'm curious to hear your opinions on the approach.
English
0
0
0
62
Steve Magness
Steve Magness@stevemagness·
Mastery doesn’t start with freedom. It starts with rules, drills, and rigidity. You ingrain the fundamentals. You build precision. But here’s the paradox: true expertise ends with play. The arc of mastery is rigid → flexible → playful.
English
4
11
77
7.5K