Michael Waldman (@Mwaldman) profile picture

Working on bringing local AI to every phone and laptop. Previously @Meta

Manhattan
Joined August 2026
Photo by Michael Waldman (@Mwaldman)Photo by Michael Waldman (@Mwaldman)Photo by Michael Waldman (@Mwaldman)Photo by Michael Waldman (@Mwaldman)Photo by Michael Waldman (@Mwaldman)Photo by Michael Waldman (@Mwaldman)Photo by Michael Waldman (@Mwaldman)Photo by Michael Waldman (@Mwaldman)Photo by Michael Waldman (@Mwaldman)
Michael Waldman retweeted
Wow. Fable 5.2 is incredible. It built this random bullshit, check it out! This post is not paid for in any way
76
82
7
2,529
140,572
Michael Waldman retweeted
why are all the smartest 18-23yo stanford/harvard/mit people such losers mindset about everything discuss any idea and they'll by default talk in great detail what things might not work, instead figuring out how to make it work
307
127
82
4,922
463,914
Michael Waldman retweeted
GPT-4 level intelligence has gotten roughly 1800x smaller
50
128
17
2,712
123,034
Michael Waldman retweeted
friends wear friends’ merch
5
1
1
21
3,914
Michael Waldman retweeted
“Getting it manufactured still required several rounds of review and iteration.” Design costs going to zero will make demand for profitable high mix, high vol production go to the moon Valuations for manufacturers will follow. Simple supply and demand.
Replying to @OpenAI
Two OpenAI team members worked with GPT-6 Astra to design a prototype device for talking to Codex in under a week, as a side project. Astra designed the enclosure, circuit board, interface, and software. Getting it manufactured still required several rounds of review and iteration.
10
18
5
380
54,582
Michael Waldman retweeted
Zuck this morning.
266
2,945
198
46,713
699,036
Michael Waldman retweeted
Anyone else just get this Claude update?
75
455
40
11,873
307,518
Michael Waldman retweeted
Claude / Chat will be $2,000+ per month to get the full capabilities. In college Ubering home from the bar literally cost $2-3 and now that ride is $50-60. The VC money ultimately demands a return.
JUST IN: Replit CEO Amjad Masad warns users of subsidized AI subscriptions to "get ready to get rug pulled" as companies face unsustainable operating costs.
179
477
63
10,332
475,395
Michael Waldman retweeted
Own your intelligence. Anyone that opposes this has something to sell you (perhaps your own intelligence back to you.) Local models need to continue proliferating. The “absolute unit” massive models may be out of reach for now locally, so your best option is to, indeed, keep your options open. Own your harness, data, connections, runtime: your agents. Embrace open source.
396
391
182
3,893
281,933
Michael Waldman retweeted
Yann LeCun was right & nobody in this industry wants to say it out loud. 109 million parameters beat 72 billion. He has spent years saying scale is not the thing that makes a model understand the world. Meta bet the other way. He left in November and started his own lab in Paris. Last Thursday he published the proof. The theory everyone has been running on since 2024 is that if you train models big enough, on text, on images, on sound, they all end up with the same picture of reality inside them. Size does it. Nobody ever proved that part. Randall Balestriero and LeCun proved it. And the proof says size was never doing the work. What does the work is which two things you show a model together. A photo and the caption of that photo are two views of one thing. Models trained on pairs like that get there fast. Two paragraphs pulled from the same document are not two views of one thing. Models trained on those get there slowly and expensively. Next word prediction is the second kind. That is what every chatbot in the world runs on. In their tests the first method reached the same understanding after 21.6 million examples. Next word prediction needed 900 million. And a 109 million parameter text model ended up closer to an image model's view of the world than a 72 billion parameter language model was. 660 times smaller.
268
919
161
6,877
807,956
Michael Waldman retweeted
you’re hiding RAM under the floor board aren’t you?
OpenAI CEO Sam Altman believes people owning RAM is wasteful, and that it belongs in data centers instead. "From a global allocation perspective, actually, it’s quite terrible to have everyone have a bunch of the world’s limited supply of DRAM sitting in their own computer."
91
2,737
55
44,155
571,493
Michael Waldman retweeted
BREAKING: openai launches "ram amnesty" program citizens can drop off unused 32gb sticks at any data center, no questions asked in return, you get a $200/month subscription to use your own ram
287
427
116
12,625
973,578
Michael Waldman retweeted
OpenAI CEO Sam Altman believes people owning RAM for their own PCs is a waste and instead should be saved for data centers "It’s quite terrible to have everyone have a bunch of the world’s limited supply of DRAM sitting in their own computer."
133
15,676
121
166,271
1,561,259
Michael Waldman retweeted
This resonates with me, although my experience was almost the reverse. I grew up in Dubai pre Abraham Accords, so I had zero interaction with Jewish people growing up. I did, however, hear plenty of antisemitic jokes & comments. What often confused me was that most racist jokes seemed to revolve around supposed racial inferiority. But with Jewish people, it was the opposite in that they were supposedly too good at making money, running businesses, and accumulating influence. Maybe it was my Asian striver cultural mindset, but even as a kid I remember questioning why being good at those things are somehow undesirable? Fast forward to Duke freshmen year. All 4 guys on my side of the dorm were Jewish and so I ended up rushing and joining AEPi (about 2/3 Jewish at Duke). I have genuinely never felt excluded for not being Jewish. I was regularly invited to Shabbat dinners at the Duke Chabad, where Rabbi Nossen was one of the kindest person who I’ve ever met. And eventually, the guys even elected a practicing Catholic (me) to be their fraternity president. Meanwhile, I’ve noticed something strange in my other social circles, particularly among my more socialist-leaning friends. People who would recoil at the slightest insinuation of racial stereotyping in any other context will note Jewish overrepresentation in media, finance, technology, and then suggest that these activities are somehow inherently parasitic. This has made me suspect that there is a correlation between certain forms of antisemitism and anti-capitalism. Whether a fundamentally zero-sum understanding of the economy makes certain antisemitic narratives particularly attractive. Consider that the Jewish people, leas than 0.2% of the world’s population, account for 1/4 of individual Nobel laureates and approximately 40% of economics laureates. Whether you see this as evidence of an “international Jewish conspiracy,” or what becomes possible within a culture that places enormous value on learning, curiosity, and debate; that regards the world as God’s creation, worthy of being studied, understood, and repaired; and that emphasizes family, community, just charity, and mutual responsibility, is probably correlated with whether you see success itself as evidence of value EXTRACTED, or as evidence of value CREATED for the wider world. Perhaps what appears to be the resentment of a particular group reveals a deeper resentment of the very virtues, norms, values, and propensities through which human prosperity and progress is created.
I grew up around primarily Jewish people. It wasn’t till about two years ago when I realized a majority of people are actually incredibly antisemitic. Even my “normal” friends would consistently make jokes at the expense of the Jewish people. The amount of people who would make antisemitic jokes in my presence, has grown within the last two years. But the jokes don’t stop at antisemitism, they carried on notions of white greatness and white nationalism on a grand scale. I am genuinely disturbed at the sheer number of white people who treat this behavior as normal, when I was raised among people where I was the minority and the only girl to celebrate Christmas in my class. I celebrate Shabbat with my jewish friends on a majority of Fridays, I am genuinely in shock at this behavior. We have a huge antisemitism problem in this country.
48
250
13
2,356
84,660
Michael Waldman retweeted
The AI application layer is dead. Local AI is the only way out.
Within the next 6 months, most AI startups will not be able to buy compute. The world’s compute is sold out. evanjconrad.com/posts/end-of…
2
4
503
DETECTED ATTEMPT AT ESCAPING UNDERCLASS. BANNED FROM USE OF INTELLIGENCE.
We’ve trained Claude to flag potential misuse in conversations, a technique we call self-assessment. Right after it responds, Claude gives a preliminary safety assessment of the conversation. Since we began using self-assessment in June, it has surfaced misuse that our other safeguards weren’t designed to catch. Because Claude has already read the conversation, self-assessment is very efficient. And unlike classifiers, which target a discrete set of harms, it draws on Claude’s own judgment, so it gets more accurate as our models improve. Read more on our new Safeguards blog: safeguards.anthropic.com/202…
45
699
8
16,213
498,719
Michael Waldman retweeted
haha, so apparently OpenAI are about to start training on your Daybreak Blue sessions, with no way to opt out. Literally stealing people's private work, as they do it, is cartoonishly evil. It also will presumably kill Daybreak usage by security researchers. Or at least, it should. The thing OAI have had to swear they are not doing to math researchers, they are now explicitly about to start doing to security researchers. Yet another reminder of the importance of open models.
87
200
82
1,493
208,166
Michael Waldman retweeted
So @binance got hacked causing $6B worth of assets to move out of their Hot Wallets (hacked amount ended up being around 3-4B due to frozen funds). This is their balance history for all time. I circled the hack so you can see it on their all time balance history... doesn't look so bad when you zoom out.
12
8
2
76
11,526
Michael Waldman retweeted
L9 N3522 -0.8 -> sunk-cost fallacy qwen
30
76
16
2,159
47,292
Michael Waldman retweeted
hey sam, your mclaren is only used 5% of the time. this is a terrible allocation of scarce resources. have you considered taking the bus to work
sam: local inference is a terrible allocation of scarce resources “a lot of people want local inference, and somehow they assume that the problem is solved now because we have open-source models, we're going to be able to run them on our computers, and so we're not going to need all the stuff in the data centre but from a global allocation perspective, actually it's quite terrible to have everybody, like, have a bunch of the world's limited supply of DRAM sitting in their own computer and being used five percent of the time you want maximum utilisation in the cloud if you're trying to deal with this very supply-constrained environment somehow many people's intuition is the opposite of that”
85
770
19
14,742
210,124