An introvert who doesn't like to be alone. CTO @udaandotcom
Bangalore, India
Joined May 2007
- Tweets6.2K
- Following517
- Followers2.4K
- Likes2.9K
Siddhartha Reddy retweeted
There's some senior engineers on here who, in the face of AI changing the very structures this industry is built on, present themselves as Stewards of the Craft of Programming.
Bugs the hell out for me.
They keep alluding to the virtue of doing things by hand, the value of Good Code and of craftsmanship, saying that yes, AI might change some things, but hey it won't ever change the value of something done with proper care, by hand, just like in the olden days of 2024. Most everything that was Good back then is still Good now: deep knowledge of languages and systems, good code, great architecture. You know the sermon.
And, yes, some of that *is* true. At least I think so. But we don't know, because no one truly knows what will be important in the future.
We're going through times of great change and I think the best advice to give to young people and others who want to have a job and provide value is: keep your eyes open, rethink from first principles, derive from what businesses value what you can provide; the digital world is more malleable than ever, you should rethink how you can help shape it.
All this pounding on the Bible of Good Code seems to me to be doing the opposite: it lets people feel comfortable -- what I've valued all my life is still valuable, I can rest easy -- in times when one shouldn't feel comfortable. This is a bumpy, exciting, strange ride. If you lean back you might fall off the waggon.
Let's not delude ourselves and others that we can keep doing what we've been doing. Rethink it all.
And, like I said, it bugs the hell out of me seeing senior engineers (that are secure in their career) talking as if there's nothing *really* to change, because the Fundamentals will still be valued. They might!
But it's the wrong advice to not rethink everything right now.
Siddhartha Reddy retweeted
>be me
>discover effective altruism
>apparently normal charity is inefficient
>why donate to random sad thing when spreadsheet can tell you optimal sad thing
>fair enough
>buy mosquito nets
>save lives
>numbers look good
>feel powerful
>couple years later
>someone asks an innocent question
>why only count people alive today
>huh
>future people matter too
>obviously
>my grandchildren shouldn't matter less just because they haven't spawned yet
>reasonable.jpg
>keep following logic
>what about their grandchildren
>also yes
>what about people in 500 years
>sure
>5000 years
>why not
>500 million years
>starting to get weird but morality is morality
>open calculator
>humanity could survive for an astronomically long time
>could colonize galaxy
>could have trillions upon trillions of descendants
>maybe digital people too
>maybe simulated civilizations
>maybe dyson spheres full of happy uploaded minds
>calculator starts smoking
>realize currently living humans are rounding error
>8 billion people suddenly looking extremely beta
>future contains potentially 10^something people
>can't even fit beneficiaries in google sheets
>new moral priority unlocked
>protect the long-term future
>stop thinking in units of "people helped"
>start thinking in "fraction of cosmic endowment preserved"
>malaria?
>terrible
>but only kills existing humans
>AI extinction could delete the entire light cone
>nuclear war could permanently derail civilization
>bad institutions could lock in terrible values for ten million years
>someone invents wrong constitution in 2140
>quadrillions suffer
>better fund governance workshop now
>friend says maybe we should improve hospitals
>explain opportunity cost
>friend says hospitals are full of actual sick people
>explain scope sensitivity
>friend stops inviting me to dinner
>need to decide what to fund
>easy
>expected value
>suppose project has one in a million chance of preventing extinction
>sounds tiny
>but extinction destroys 10^50 future lives
>multiply
>mother of god
>$10 million project has expected value of several galaxies
>charity evaluation complete
>someone asks where the one-in-a-million number came from
>expert judgement
>which expert
>us
>how calibrated
>extremely thoughtfully
>reduce estimate to one in ten million to be conservative
>still beats curing cancer by 38 orders of magnitude
>epistemic robustness achieved
>someone says maybe project doesn't work
>assign 20% chance
>still astronomical
>maybe project makes problem worse
>assign 5% chance
>still astronomical
>why 5
>because 30 felt pessimistic
>publish 46-page report
>contains seventeen sensitivity analyses
>every sensitivity analysis begins after assuming intervention has positive sign
>critic says you're multiplying enormous hypothetical stakes by extremely uncertain probabilities
>yes
>that's literally why it's important
>critic says the uncertainty might be structural rather than numerical
>make probability smaller
>critic says no, I mean maybe your model is wrong
>make probability smaller again
>critic begins rubbing temples
>discover AI safety
>perfect longtermist cause
>AI might kill everyone
>or create utopia
>or seize galaxy
>or tile universe with paperclips
>or create billions of conscious software minds
>finally a problem with numbers big enough for me
>start AI safety nonprofit
>mission: prevent dangerous AI
>hire smartest people available
>smartest people immediately start building better AI to understand dangerous AI
>interesting
>we must understand capabilities to understand safety
>we must scale models to study alignment
>we must race ahead so less responsible actors don't get there first
>we must deploy systems to learn how deployment can go wrong
>we must build the thing quickly because building the thing quickly is dangerous
>outsider asks why the people most worried about AI apocalypse all work at AI companies
>complicated field
>company releases stronger model
>very concerned
>company begins training even stronger model
>extremely concerned
>company raises $14 billion
>concern reaches unprecedented levels
>need to influence government
>future is at stake
>normal democratic process too slow
>politicians don't understand exponential curves
>public doesn't understand x-risk
>experts must guide them
>who counts as expert
>people who understand x-risk
>who understands x-risk
>our friends
>someone objects that this seems politically convenient
>explain we're representing future generations
>future generations unavailable for comment
>develop concept of value lock-in
>terrifying possibility that one ideology controls civilization forever
>therefore extremely important that civilization adopts correct values before lock-in
>whose values
>let's circle back
>begin with impartial morality
>end with small group of people deciding what quadrillions of hypothetical beings would want
>beautiful arc
>meanwhile actual humans keep doing annoying things
>voting wrong
>having parochial attachments
>loving family more than strangers
>caring about local community
>getting upset when told their suffering is cosmically negligible
>evolutionary biases everywhere
>explain that moral intuition cannot be trusted
>except intuition that future digital people count
>and intuition that extinction is uniquely bad
>and intuition that our probability estimates are sane
>and intuition that our institutional choices improve the future
>those intuitions survived peer review
>someone donates $5k to local homeless shelter
>inefficient
>could have funded 0.0000000000003% of an AI governance researcher
>think of all the simulated people you just killed
>okay maybe don't phrase it that way publicly
>PR team says "future generations deserve a voice"
>much better
>journalist asks what longtermism means
>say "future people matter"
>everyone agrees
>great
>journalist asks what follows from that
>well technically we should redirect enormous resources toward low-probability interventions affecting astronomical futures
>journalist raises eyebrow
>return to "future people matter"
>motte has entered the chat
>critic: of course future people matter
>me: glad we agree
>critic: I don't agree that your institute knows how to help them
>me: why do you hate our grandchildren
>eventually notice uncomfortable implication
>if future value dominates everything
>then helping people today mostly matters through effects on future
>education matters because future institutions
>health matters because future productivity
>democracy matters because future trajectory
>human beings slowly become instrumental variables in their own moral philosophy
>see starving child
>feel compassion
>check spreadsheet
>child's direct welfare contribution negligible
>but perhaps childhood nutrition improves national institutional quality
>compassion restored
>tell myself this is impartial altruism
>one day assistant asks obvious question
>"how do you know your intervention actually improves the far future?"
>silence
>open spreadsheet
>increase column width
>add confidence interval
>assistant asks again
>"no, I mean how do you know the sign is positive?"
>stare into cosmic light cone
>10^50 people staring back
>none of them exist
>none of them can tell me
>none of them can falsify my assumptions
>realize I have invented the perfect constituency
>infinitely important
>completely silent
>and always represented by me
Siddhartha Reddy retweeted
The "whiteboard defense:" I should be able to pull you aside at any moment and ask you to explain any customer-facing system you've shipped. You should be able to clearly explain how it works and defend the decisions you made. This is my benchmark for responsible AI usage.
I don't expect line-level familiarity with the code. I don't care if you remember the exact function name or implementation detail. You may not even know it. I don't care.
But if I ask "why did you do X instead of Y?", "what happens if this actor behaves maliciously?", "what data structure did you use here and why?", or "where does this fail?" you should be able to answer confidently.
For PoCs, demos, experiments, whatever: I don't care. Generate 100% of it and understand none of it. Speed over quality every time in those specific scenarios.
But if you're shipping customer-facing work, you can't be shipping things you don't understand at a high level.
Siddhartha Reddy retweeted
the general public has been getting constant waves of fear and confusion from ai companies
i thought the whole "you will be replaced" thing was as bad as it gets but now it's "you will literally die"
the defense is always "it would be irresponsible of me not to warn people of the risks"
but all that's happening here is the same thing that always happens: a group of people had the thought that their cause is righteous and moral
and they're incrementally justifying worse and worse behavior and entirely incapable of reflecting on how they might be the problem
Siddhartha Reddy retweeted
More than ever, we should ignore how we *feel*. "I'm worried" is a state of mind, not an objective reality.
I believe ex-frontier people *feel* the way they claim they do. But it's not whistleblowing until you leak a spreadsheet, an experiment postmortem, or some form of fact.
"I'm worried" could come from (a) real threat, or (b) "I've been grinding for years, sleeping on a mattress under my desk, and my only contact with human beings is the company Slack". In which case, you deserve a hug and a break, but not to cause mass hysteria.
Again, the threat might be real. But for us to take you seriously, you need to do two things:
1. Leak data, not feelings
2. Detach yourself from financial incentives (i.e. equity)
Thanks for coming to my ted talk, I love all human beings, including the ones under great amount of stress and pressure to perform in Silicon Valley.
Siddhartha Reddy retweeted
Pre-AI: "I can build the core feature in a week, the product around that feature in a quarter, and the company around that in a year."
Post-AI: "I can build the core feature in a day, the product around that feature in a quarter, and the company around that in a year."
So cool!
We have a huge news to share today!
Today we are unveiling the first truly accessible RL robot - welcome Microduck
A 25 cm tiny open-source biped with 15 actuators and packed with sensors (camera, speaker, LiDAR, NFC, bluetooth, wifi, etc) that you train yourself with reinforcement learning.
It's also playable out of the box with more than half a dozen fun and playful pre-trained policies to have it walk, sit, crouch, roller-skate, pick up objects with its articulated beak, and recover on its own.
And all for less than $400.
See all the details, play with the simulator and order it at: pollen-robotics.com/microduc…
(video with sound on 🔊)
Siddhartha Reddy retweeted
Me: "Please fully remove this feature."
Claude: "Done! I have removed the feature, and added some tests to ensure that the feature continues to not exist."
“ The key idea here is that the best way to be sycophantic to smart people is to disagree with them without making them feel stupid. Ideally you’ll come up with a counter-argument that works against what they’ve said but is straightforward for them to knock down by clarifying their idea. If you do it right, you’ll validate their self-image as a smart person who appreciates rigorous critique. But if you actually come up with a devastatingly rigorous critique, they won’t enjoy it at all. At best, they’ll resentfully agree with you. At worst, they’ll double down on being right and convince themselves you’re a rude idiot”
This seems so obvious once it’s pointed out that I feel stupid for not having realised it.
FWIW, I think Claude models do this far more covertly than the GPT models. This might be one reason why I dislike working with Claude in general.
Siddhartha Reddy retweeted
Another hot take: you are not the bottleneck, you are an essential part of the agent backpressure. Never ship at speeds greater than what you can review.
Unreasonably excited about this!
I've started a new company: @superlogical! We're going to begin by building a terminal multiplexer. The entire vision is much larger, but the multiplexer is the foundation. Sign up for the newsletter to get beta access and devlogs (product updates only I promise).
superlogical.com/
Siddhartha Reddy retweeted
What worries me most about protectionist arguments against Chinese AI is that they seem to ignore history.
- Groups of very smart people, with constraints (no easy access to the best nvidia chips) will out-innovate teams who can merely solve their problems with money.
- Groups of smart people who borrow every good idea they see will out-innovate those that are insular and think they're smarter than everyone else.
Some greatest American moments from our history come out of these two points. We shamelessly adopted great ideas and incorporated them, when he had limitations placed on us we got creative. And we succeeded.
In the long term, limiting Chinese access to the best nvidia chips will create more innovation, not less. Further, using distillation to explain how DeepSeek, Moonshot, et al. are delivering strong models dismisses the genuine innovations taking place.
Hunk is the best tui diff tool I’ve tried. hunk.dev/
I use it everyday. I have it bound to `prefix+d` in a popup in herdr. I’ve also set it as the pager in lazygit (which is bound to `prefix+z` in herdr).
Siddhartha Reddy retweeted
Yes -- I've been meaning to write something like this for a while.
I think a lot of people expect superintelligence to be like some magic power. A superintelligence walks into a room, blows a puff of air in the right direction to cause a complex chain reaction that makes an arbitrary thing happen.
But raw intelligence is not actually *that* powerful. It's just a useful tool.
I think the rationalist AI doomer crowd worships intelligence in a sort of religious way and that's why they think it's so dangerous, but... it's not really.
If you lock a superintelligence in a room with an angry grizzly bear, the grizzly bear is gonna win.
If raw intelligence were all that mattered, we wouldn't have our current president.
etc.
