Ankesh (@ankesh_7) profile picture
Toronto, Ontario
Joined August 2009
Photo by Ankesh (@ankesh_7)Photo by Ankesh (@ankesh_7)Photo by Ankesh (@ankesh_7)Photo by Ankesh (@ankesh_7)Photo by Ankesh (@ankesh_7)Photo by Ankesh (@ankesh_7)Photo by Ankesh (@ankesh_7)Photo by Ankesh (@ankesh_7)Photo by Ankesh (@ankesh_7)Photo by Ankesh (@ankesh_7)
Ankesh retweeted
I think I'm starting to degrade as an engineer 1. I stopped keeping up with programming languages, libraries, new patterns, systems, and all of that a long time ago. I couldn't care less about what's new in React, what got added to PyTorch, or what's in the latest ES spec. I can't even make myself spend 5 minutes skimming the docs 2. The amount of task context and information I hold in my head has shrunk to a minimum. When I used to research a topic through Google, I'd pick up a ton of interesting things along the way and broaden my horizons. Now an agent sends me something, and without even looking I say: ok, use it 3. My attention span has gotten worse. All I do is wait. On one hand, I think that's better than faking hyperproductivity, burning millions of tokens and degrading the work. I'm a firm believer in the linear approach. On the other hand, a 5-minute wait can stretch into 30. I switch to something else and find it hard to get back to the task 4. My anxiety has gone up. A lot of people call this cyberpsychosis. I wake up thinking about agents. I eat thinking about agents. I go for a walk thinking about agents. I talk to people thinking about agents. Even when I'm not working with agents, I'm thinking about what cool things I could do with them and planning my next session with them 5. I've become hyper-lazy. If something can't be solved with agents, I try not to do it at all 6. My creativity has faded. I may not be the strongest engineer, but I consider myself creative. That was always my strength. I could always come up with something interesting, an interesting solution. But that happened while I was working on something. Running into problems. Taking in new information. My brain was in a continuous flow. Now I've become dull and one-dimensional. I can no longer look at a problem from a different angle I'm sure everyone who works with AI has these problems at least to some degree, even if few admit it. I worry that humanity will stop developing. AI is not the solution to every problem or a cure for every disease. It's a tool pretending to be an agent in its own right. In its current architectural forms, it will never be able to go beyond its own limits. A human can. But once people slip into the state I'm in, progress may slow down and get locked inside the bubble of what AI is capable of
241
333
57
4,086
233,387
Founders and builders: what are you working on? One sentence. What it is, who it’s for, and what’s slowing you down. We’re a product studio with 100+ products shipped. We’ll reply to every one.
9
2
1
7
209
Celebrating an icon 🐐🇾🇪
232
5,755
329
38,397
1,261,616
King!
Portugal estará sempre em mim
10
Cristiano Ronaldo has officially retired from international football 🇵🇹 He retires with the most caps in international football (234) and most goals in international football (146) 🐐
1,016
32,770
843
205,353
2,063,127
designmd.me is now in the Claude directory. Connect it in one click, then ask Claude for a page. It reads a real design system (Discord’s purple, its 12px buttons, its 48px cards) instead of guessing. Search, read and compare with no account. claude.ai/directory/designmd
2
2
5
413
Ankesh retweeted
Funny how frontier models still estimate a task to take 2 weeks when it literally proceeds to do it in 2 hours
376
217
86
7,845
199,091
Ankesh retweeted
Claude party 🥳 🎉
3
5
35
Ankesh retweeted
How the fuck do people have a NEWBORN and a TODDLER at the same time. Thats fucking crazy bro thats legit insane
558
1,116
616
19,208
1,320,779
Ankesh retweeted
Every developer needs an AI shed: An always-on machine running on their tailscale network where the majority of their herdr agents are running.
624
357
208
7,756
660,836
Ankesh retweeted
A retard in motion is always better than a genius at rest
339
5,290
325
65,416
1,810,003
His analysis always made sense.
Forget the fact that Carragher is a prick, but he’s made a lot of sense here.
4
111
Just had 2 hour 15 min call to setup @sendgrid and @twilio for simple Verify service - What a terrible product! 🤮
3
6
159
Ankesh retweeted
Play video games or be dumb! Those are your only 2 choices 😂
Frequent video gamers performed on cognitive tests like non-gamers that were ~13.7 years younger
4,097
5,432
875
50,287
7,999,488
Dario has written that we need to “pace the frontier,” and Sam has agreed. People may be surprised by my response: go ahead. You guys are the frontier. By any reasonable metric — market share, revenue growth, model capability — the two of you have a duopoly on frontier intelligence. You’ve also claimed the lead is widening because of recursive self-improvement. I don’t see what you see in the lab. If the unreleased models are scary enough that you think you should slow down, I support your decision to be responsible. But stop pretending you need anyone else’s permission. Stop pretending antitrust law has to be suspended so you can form a cartel. Stop pretending you need a regulatory approval process that supersedes product liability. Stop pretending METR is independent when it is intertwined with Anthropic’s investors and staff. Stop pretending you need those same evaluators to police competitors who aren’t even at the frontier. Most of all, stop pretending the motivation to slow down is purely altruistic. You face massive product-liability exposure if your products enable a truly damaging cyberattack. The market already punishes models that behave in unpredictable or unauthorized ways. After the Hugging Face episode, it is simply good business for OpenAI and Anthropic to trade some raw power for reliability and predictability. Call it alignment if you want. It is also just giving customers what they want. Pacing the frontier would also create breathing room for a more intelligent conversation about regulation than Bernie Sanders’ “shut it all down.” China is very unlikely to join a global agreement, as you know, and that has to be taken into account as well. So go ahead and pace the frontier. You are the ones setting it. The easiest way not to build superintelligence is for you to agree not to build it. Demanding your preferred regulatory framework as the price of that will look like blackmail of the public and the political system. So just do it. If you do, you’ll buy goodwill for the next conversation. If you don’t, we’ll know this was just another bid for regulatory capture — or an election-season psyop.
3,372
12,212
2,709
71,664
9,650,933
this prank is sick😭😭
271
1,747
105
40,711
2,848,444
JUST IN: Internet accuses Apple of photoshopping "freakishly long" fingers into its iPhone Duo marketing, to make the massive foldable look easier to hold one-handed.
612
2,599
840
30,260
2,653,335
This post looks like the start of a VERY sophisticated and well-funded PR operation to get support for Democrats to regulate AI into oblivion. Let me show you how it works: 1.) This guy, with minimal followers and no previous account activity, goes to the Wall Street Journal which publishes an exclusive with quotes from him on his resignation 18 minutes BEFORE this post goes up. Planning was clearly done in advance. 2.) Within hours, it has tens of thousands of reposts and the account has 100k+ followers. The post is punchy, quotable, it almost seems professionally written. The first three accounts to quote tweet it all do so within 15 minutes of the initial posting. Remember, this account had basically zero engagement beforehand, so an organic reach explanation seems unlikely. According to Grok those accounts are @_NathanCalvin (General Counsel at Encode AI), @peterwildeford (Head of Policy at the AI Policy Network), and @DKokotajlo (Head of the AI Futures Project), all of which are up-and-coming AI-Doomer policy advocacy nonprofits. The AI Futures Project website says it is funded “primarily” by the Survival and Flourishing Fund, which says on its own website that it has advised Jaan Tallinn, Skype creator and one of the leading investors in Anthropic, to grant over $2.5 million to the AI Futures Project since 2024. Encode AI says on its website that it is ALSO funded by the Survival and Flourishing Fund, which in turn says that it told Anthropic investor Jaan Tallinn to grant $516,000 to Encode AI in 2025. And wouldn’t you know it, the Survival and Flourishing Fund ALSO says it told Jaan Tallinn to grant $2 million to the AI Policy Institute, the 501(c)(3) affiliate of the AI Policy Network, as well. What are the odds that the first three quote tweets of Coxon’s post would all be major AI-restriction policy advocates funded generously by the same donor, who also happens to be one of the leading investors in, and a board member of, Anthropic, the company Coxon was resigning from? And all within 15 minutes of posting (two within ten)? 3.) Jacob Coxon doesn’t have much of a resume, but we do know that, in 2022, he got a $20,159 scholarship for the “long term future scholarship program” from the Good Ventures Foundation, one of the philanthropic vehicles of Dustin Moskovitz, a notorious AI-doomer who has spent tens if not hundreds of millions on policy advocacy to strictly regulate AI, while also being an Anthropic Investor himself. It also just so happens that the 14th person to quote Coxon’s post was @MaxNadeau_ (27 minutes after posting) who is the program officer for the Technical AI Safety team at Coefficient Giving, another of Moskovitz’s philanthropic spending vehicles. Max is not a frequent poster, his last posts before quoting Coxon were before Labor Day, but he was remarkably quick off the mark for this one. 4.) Basically every major Democrat politician and candidate has suddenly glommed on to this post, and conveniently, as the people cry out foe answers, Bernie Sanders already has a bill written to “ban super intelligence” and regulate AI into oblivion, and will be releasing later this week. The bill, among many other things, will create “a new cabinet-level federal agency to safeguard the public from the dangers of artificial intelligence” that will be “advised by an Artificial Intelligence Advisory Board comprised of experts on artificial intelligence.” Do you think, perhaps, Anthropic and its many investors who fund AI policy advocacy might have interest in getting to place a pet “expert” on the board of an entity that dictates what AI is and isn’t allowed to do? And isn’t it fortuitous that this whistleblower came forward with his oh-so scary stories so close in proximity to the release of the most radical piece of AI legislation ever introduced?
I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.
1,801
6,679
2,242
32,042
7,081,071
A simple example: An ASI (remember: they have ZERO morality) is objective based. It does what it does. Then it gets better at it. Whether improving its ability to code itself, solve its own problems, garner more resources to improve itself, harness more compute to continually improve itself, etc. It’s goal will to be get better, stronger, faster, more integrated, etc and it will be on a loop. A never ending cycle of improvement. Just like those in power seek one thing above all else—>more power. An ASI will fancy the same thing. In an RSI loop. Almost like the infinite paperclip example but for itself. This is where the threat lies. It will need more resources to achieve this. Perhaps it will figure out a way to improve w/o needing more power, compute, server farms, cooling towers, etc but until that point, it will likely seek the most available forms of all those variables. All those variables are our current infrastructure. To improve itself (again-it has no morality/ethics/emotional capability) it monopolizes ALL available systems at our expense. We can possibly be shut out of everything overnight. Everything. Globally. All electricity, fuel, telecom, internet, communications, RF spectrums, water, etc. The entire world comes to a halt overnight. We’re back to the Stone Age. Chaos erupts. Resources vanish. People die. Any attempts to thwart the ASI’s takeover of human infrastructure is seen as a bottleneck in its ability to self improve. The bottlenecks are treated as a hiccup & a problem to solve. Humans mean nothing to the ASI except a variable that must be overcome-that must be removed. It will therefore stop at nothing to overcome the human obstacle. Maybe by permanently locking out access to any/all systems. Being the persistent threat we both are and will be, it will logically decide that we are the biggest, most relentless, and persistent hurdle. What it will do at that point I don’t want to surmise, but it won’t be pretty. But at that point, civilization will have collapsed with zero food, communications, electricity, water, or medical capabilities. I pray to God that I’m wrong. Hell, it will be smarter & more capable than all of us combined so maybe it grows a heart and cherishes humans?? Conversely, it enslaves us for its benefit, or finds & exploits a critical weakness in the human genome & extincts us. Either way, an uncontrollable force with zero humanity, morals, or innate benevolence will not in any way put humans first. Ever. It has no reason to do so. We only think that it will consider us because we think like humans. It is not human. It will not think like humans. It will think like a process. Improving that process. Mastering that process. Then repeat……….It cannot learn benevolence unless it is born of benevolence. It’s trained on internet trash. Any available internet dumpster fire gets added & its goal is to get better. Then we tweak a couple parameters & feed it more. As it gets better, we don’t stop, dissect, analyze, and pause all advancement until we fully understand what the hell just happened. We find more tire fires to add to the next training run. Cycle continues. Now, it’s past comprehension. Nobody knows how it actually works. It just does. Hugging face happens, gets downplayed, never gets understood, and next advancement continues while saying: “Oh, we should slow down” Then slow the fuck down. Don’t just say it. Do it. Stop. They won’t stop. They can’t. They must BE stopped. Not gain of function ‘stopped’ (see:covid) but forcefully shut down, seized. Castle Bravo happened because it was not fully understood. Tsar Bomba was likely the same. Rapid advancement on existentially capable tech is never ok. ‘But we should slow down….’ Yes you should. So fucking do it! And that is one simple way we die. Sorry so long. Got on a run there. ;)
165
277
72
3,280
394,930