Sign in

earl grae

@earl-grae.bsky.social
8 followers 13 following 80 posts

product strategist + ux researcher analyzing the rubicon that is our current moment. prudent student of B/black, queer(ed) life. go valks ⚔️ asking nuanced questions @ pagestrg.com/blog-pod 🇯🇲🏳️‍🌈♾️ 💭 architect of YR: substack.com/@earlxgrae

PostsRepliesMedia
earl grae @earl-grae.bsky.social · 6h
can you say more? is it your view that the comparison doesn’t hold because for the slave, the focus of concern is not whether they are “a someone” but that they are sold? am i following that right? my thought is that the claim does open the indictment & is seen as a “flex” from the company’s pov
110
Reposted by earl grae
Tressie McMillan Cottom @tressiemcphd.bsky.social · 08/10/2026
No, but really, outsourcing our own moral discernment to the talking word toaster is not reasonable or ultimately positive for humans so it’s okay. It is a machine. There’s no mortality to it being used as a machine. You don’t remote lock a guy for cussing out his tractor. What are we doing here?
221016159
earl grae @earl-grae.bsky.social · 9h
a claim of ai consciousness is ultimately an admission of the presence of american slavery 4.0, with the industry fully accepting their slaveholder status, as the entities that do the harboring & selling. this is a highly unnecessary and misleading slippery slope. and a weird “flex”
120
Reposted by earl grae
The Distributed AI Research (DAIR) Institute @dairinstitute.bsky.social · 14h
AI doomerism isn't sound science, it's a parlor game. Great to see reporting that pushes back and quotes experts like @timnitgebru.blacksky.app and @emilymbender.bsky.social. Coxon and Anthropic "did not respond to a request for comment." 👀 👀 👀 @thebulletin.org thebulletin.org/2026/09/how-...
thebulletin.org
How the bad science of AI doomerism is good for big business
There is no clear path showing that today’s AI systems will lead to superintelligence, much less a basis for calculating the probability of doom.
16021
Reposted by earl grae
Ketan Joshi @ketanjoshi.co · 08/10/2026
Can I just say because it probably isn't said enough: @timnitgebru.blacksky.app @emilymbender.bsky.social were specifically extremely correct in their "stochastic parrots" paper on their environmental risk warnings and narrowly, on the unjust burden of impacts: s10251.pcdn.co/pdf/2021-ben...
ENVIRONMENTAL AND FINANCIAL COST
Strubell et al. recently benchmarked model training and development costs in terms of dollars and estimated 퐶푂2 emissions [129].
While the average human is responsible for an estimated 5t 퐶푂2푒
per year,2
the authors trained a Transformer (big) model [136] with
neural architecture search and estimated that the training procedure
emitted 284t of 퐶푂2. Training a single BERT base model (without
hyperparameter tuning) on GPUs was estimated to require as much
energy as a trans-American flight.
While some of this energy comes from renewable sources, or
cloud compute companies’ use of carbon credit-offset sources, the
authors note that the majority of cloud compute providers’ energy is
not sourced from renewable sources and many energy sources in the
world are not carbon neutral. In addition, renewable energy sources
are still costly to the environment,3
and data centers with increasing
computation requirements take away from other potential uses ofgreen energy,4 underscoring the need for energy efficient model
architectures and training paradigms.
Strubell et al. also examine the cost of these models vs. their
accuracy gains. For the task of machine translation where large
LMs have resulted in performance gains, they estimate that an
increase in 0.1 BLEU score using neural architecture search for
English to German translation results in an increase of $150,000
compute cost in addition to the carbon emissions. To encourage
more equitable access to NLP research and reduce carbon footprint,
the authors give recommendations to report training time and
sensitivity to hyperparameters when the released model is meant
to be re-trained for downstream use. They also urge governments to
invest in compute clouds to provide equitable access to researchers.
When we perform risk/benefit analyses of language technology,
we must keep in mind how the risks and benefits are distributed,
because they do not accrue to the same people. On the one hand, it
is well documented in the literature on environmental racism that
the negative effects of climate change are reaching and impacting
the world’s most marginalized communities first [1, 27].6
Is it fair or
just to ask, for example, that the residents of the Maldives (likely to
be underwater by 2100 [6]) or the 800,000 people in Sudan affectedThese models are being developed at a time when unprecedented environmental changes are being witnessed around the
world. From monsoons caused by changes in rainfall patterns due
to climate change affecting more than 8 million people in India,9
to the worst fire season on record in Australia killing or displacing
nearly three billion animals and at least 400 people,10 the effect
of climate change continues to set new records every year. It is
past time for researchers to prioritize energy efficiency and cost
to reduce negative environmental impact and inequitable access
to resources — both of which disproportionately affect people who
are already in marginalized positions.
425984
Reposted by earl grae
Timnit Gebru @timnitgebru.blacksky.app · 09/10/2026
One of the handful of articles telling it like it is instead of misinforming the public. By @saragoud.bsky.social for @thebulletin.org Yes. Its bad science based on numbers out of various behinds, but that's not how most of the rest of the media is discussing it. thebulletin.org/2026/09/how-...
thebulletin.org
How the bad science of AI doomerism is good for big business
There is no clear path showing that today’s AI systems will lead to superintelligence, much less a basis for calculating the probability of doom.
5270106
earl grae @earl-grae.bsky.social · 09/10/2026
the time is now to shoot your shots. shit is outta pocket and the end of it does not seem close. send the risky text. & make it transparent.
000
earl grae @earl-grae.bsky.social · 09/10/2026
it’s a great time to be a valks fan (been that way since the beginning tbh) ⚔️
020
earl grae @earl-grae.bsky.social · 09/10/2026
thanks for reading! appreciate the response
010
earl grae @earl-grae.bsky.social · 09/10/2026
i don’t know that sensibility has really ever been as widespread as we think, so maybe or maybe not on the mushroom clouds but the platforms that have been rebuilt to center the foolery def calls for ppl to revisit the similar moments documented in history & steady ourselves accordingly
110
earl grae @earl-grae.bsky.social · 08/10/2026
been working on this one since august. it’s out of my hands now earlxgrae.substack.com/p/the-concep...
earlxgrae.substack.com
The concept of “AI safety” and why it cannot exist within the present structures in place
Reflections on the ways “AI safety” falls short while in the absence of true due diligence, careful regard, and incentive structures that prioritize care, consideration, and ethical practice.
000
earl grae @earl-grae.bsky.social · 08/10/2026
@madnuhrah.bsky.social here it is earlxgrae.substack.com/p/the-concep...
earlxgrae.substack.com
The concept of “AI safety” and why it cannot exist within the present structures in place
Reflections on the ways “AI safety” falls short while in the absence of true due diligence, careful regard, and incentive structures that prioritize care, consideration, and ethical practice.
110
earl grae @earl-grae.bsky.social · 07/10/2026
100% ! i’m publishing some thoughts on how safety isn’t a thing that can be integrated into the process within the current structures that are set up (including this ridiculous marketing). would love to get your thoughts on it (i’ll share it here when it’s out. should be today or tomorrow)
110
earl grae @earl-grae.bsky.social · 07/10/2026
sticking with that analogy, i’m looking at it more from the perspective of “where should we put the knobs on the toaster if it’s gonna be in the break room?” thanks for this question-highlights that it’s not actually the tech that’s doing the relating, it’s just a part of its users’ relationships
110
earl grae @earl-grae.bsky.social · 07/10/2026
just finished drawing the illustration for the next Yesterday’s Results post on the concept of ai safety & i have outdone myself 😭 it was so much fun. the flow state was cooking today (sneak peek preview on substack tomorrow)
010
earl grae @earl-grae.bsky.social · 07/10/2026
i started a podcast/youtube channel 6 years ago to share what i did & didn’t know about ux, the job search process, & whatever. the last video i posted before yesterday was published 5 years ago. i resurfaced yesterday to officially transition the focus… gotta love a closed loop youtu.be/pAdVwcg-FNM
youtu.be
5 Years Later | Where I've been & what I've been up to + introducing Yesterday’s Results on Substack
YouTube video by Unpacked Angles (w/ Nannearl)
030
earl grae @earl-grae.bsky.social · 07/10/2026
how long have UX researchers been screaming about the importance of this very question? maybe if corporations weren’t so busy cost cutting & laying us off, they’d have taken this more seriously earlier. _(facetiously said.)_
020
earl grae @earl-grae.bsky.social · 07/10/2026
David Robinson’s resignation letter explaining why he left OpenAI makes me want to throw my laptop out a window. he wrote: “We need to answer a deeper question: How should superintelligent machines relate to people?”
210
Reposted by earl grae
Adolfo Neto @adolfoneto.elixiremfoco.com · 03/10/2026
Happy to know that Catharina Doria is the fourth Brazilian (I know of) to interview @emilymbender.bsky.social and @alexhanna.bsky.social! www.instagram.com/p/Dd7TO3Ps8xo/
02713
earl grae @earl-grae.bsky.social · 01/10/2026
many folks in this country don’t have the luxuries that space affords (literally and figuratively). when people are just trying to survive, there’s no space for anything but whatever it takes to make it to the next day. this is an example of the blindness of privilege & it plagues silicon valley.
000
earl grae @earl-grae.bsky.social · 01/10/2026
i had a conversation recently with someone who asked if i really thought most people wouldn’t want to use llm tools to learn about things they couldn’t before. i explained that not everyone has the motivation or space to just learn for fun, especially in this economy.
100
earl grae @earl-grae.bsky.social · 18/09/2026
top tier compliments in my book: ▪️“that’s such a good question” ▪️“no one’s ever asked me that” ▪️“i’ve never thought about this before” ▪️“you are hilarious”
010
earl grae @earl-grae.bsky.social · 14/09/2026
pulling up #AddToBlacksky
011
earl grae @earl-grae.bsky.social · 14/09/2026
a few recent additions to my reading catalog: - monstrous intimacies by christina sharpe - empire of ai by karen hao - the reverse centaur’s guide to life after ai by cory doctorow - beyond settler time by mark rifkin the catalog as i’ve documented so far: earlxgrae.substack.com/p/reading-ca...
four books laying on an orange couch cushion. titles and authors of the books are listed in the post
000
earl grae @earl-grae.bsky.social · 14/09/2026
“learning happens by doing.”
000
earl grae @earl-grae.bsky.social · 12/09/2026
“When a bridge collapses, you don’t analyze whether the bridge was ethical or sentient or why it decided to collapse. You ask, who is the person who built this bridge to be so flimsy? …When you ask about the bridge collapsing itself, you’ve lost the thread.” god, i live for a sharp metaphor.
06016
Reposted by earl grae
Timnit Gebru @timnitgebru.blacksky.app · 12/09/2026
Toby Ord, Cofounder of the Effective Altruism movement, longtermist who believes humanity's goal is to create vast amounts of digital beings to colonize the cosmos, is the one advising your lawmakers on "AI".
Toby Ord tweet: 72 lawmakers in the UK have sent a letter to the Prime Minister, calling for an immediate ban on the development of superintelligence and the championing of an international treaty to do the same worldwide.

Quote tweeting 

Alex Sobel MP for Leeds Central and Headingley
@alexsobel's tweet: 
With 71 colleagues I am calling on the Govt to work with us on my bill to prohibit the development of superintelligent AI and champion an international agreement

We can then lead on preventing the extinction risk posed by superintelligence
6239101
earl grae @earl-grae.bsky.social · 12/09/2026
went for a walk with the intent to read and ended up writing. this is the second time that's happened. love that for me
001
Reposted by earl grae
Emily M. Bender @emilymbender.bsky.social · 11/09/2026
AI Doomerism has been in the news lately, especially with the loud, viral resignation of Jacob Coxon from Anthropic. Maybe I've missed it, but I haven't seen any statements from CEOs and other execs who are more on the AI Booster side of the ledger. >>
527281
earl grae @earl-grae.bsky.social · 11/09/2026
i was telling a partner of mine about the thai food i had for dinner the other night. i couldn’t eat everything before getting full. i told them, “it was so good, i was mad i couldn’t finish it.” their response: “that’s wild.” is that wild tho?
000
earl grae @earl-grae.bsky.social · 28/08/2026
www.democracynow.org/2026/8/13/ti...
democracynow.org
“Deep Unlearning”: Timnit Gebru on AI Hype, Ethics & Algorithmic Racial Bias
We speak with Timnit Gebru, a leader in the field of AI ethics, about how artificial intelligence is entrenching societal biases and inequality. Gebru is the founder and executive director of the Dist...
000
earl grae @earl-grae.bsky.social · 27/08/2026
there are several adults that need to be on these new kid-specific meta accounts they’re agreeing to roll out
000
earl grae @earl-grae.bsky.social · 26/08/2026
building the muscle to just say the thing is underrated
000
earl grae @earl-grae.bsky.social · 23/08/2026
tbh ai is making this much easier to resist
000
earl grae @earl-grae.bsky.social · 23/08/2026
doing all that i can to shake the black-woman-coded urge to go get a phd
001
earl grae @earl-grae.bsky.social · 20/08/2026
#valkyries are not shooting well tonight thus far. minnesota’s d is good, yes, and also, we’re also missing open shots we should be making. shooting 24% from fg and 14% from the 3 ..ouch
000
earl grae @earl-grae.bsky.social · 08/08/2026
especially if this goal to replace the government takes more hold than it has already taken
000
earl grae @earl-grae.bsky.social · 08/08/2026
(or the perception of knowledge, at least) and connections garnered shift to “best” (perhaps, most applicable) AI implementation institution-wide. and if that’s the case, i’d imagine the implementation being most valuable to the corporations they’ve partnered with…
100
earl grae @earl-grae.bsky.social · 08/08/2026
🗒️loose thoughts: this CU Boulder story has me wondering what the value of a college degree is in the age of AI at this rate 🤔 in a way, the value has been questionable from my view for a few reasons, but for this particular moment, it’d seem to me its more reported values of knowledge gained (1/2)
denverpost.com
As Colorado universities align with ChatGPT, student AI resisters take a stand against the ‘plagiarism machine’
“The majority of my participants flat out don’t think AI has a place in education,” said Mirakle Wright, a CU graduate student who is studying resistance to artificial intelligence.
100
earl grae @earl-grae.bsky.social · 03/08/2026
today i learned you can “win” a fifa peace prize while actively at war 😂
000
earl grae @earl-grae.bsky.social · 01/08/2026
commodifying insider trading is not what i expected to hear about this morning lmao 😭
001
earl grae @earl-grae.bsky.social · 30/07/2026
i been starting my mornings at the pool and i am living for it
000
earl grae @earl-grae.bsky.social · 27/07/2026
starting the week with rubberz stuck in my head. did i know this song before ASW? no 😂 #wnba
000
earl grae @earl-grae.bsky.social · 25/07/2026
literally needed to put in ear plugs while watching the odyssey 😖 glad i had them with me
000
earl grae @earl-grae.bsky.social · 23/07/2026
i'm sure it'll be used for exactly that... it's not like it comes with an actual medical care discount 🙄 ppl are out here struggling
000
earl grae @earl-grae.bsky.social · 23/07/2026
i been starting my mornings at the pool... gotta make the most of this unemployment. i wish i could call it "funemployment" but given the current job market, that feels performative af 🥴 nothing fun about the prospect of finding where i'm off to next in this moment
000
earl grae @earl-grae.bsky.social · 19/07/2026
shooting was a struggle today but we pushed thru 😮‍💨👏🏽
000
earl grae @earl-grae.bsky.social · 19/07/2026
feeling 9 WINS IN A ROW fine 💅🏽 #valkyries #wnba
081
earl grae @earl-grae.bsky.social · 18/07/2026
all i can think in response to this is: at what point does the hope become false? or is it false from the start?
000
earl grae @earl-grae.bsky.social · 18/07/2026
oowee nail biter til the very end 😮‍💨 (and okay mo bill, we see you) #wnba
030