خیلی دلم میخواهد بدانم ایرانیهای آلبرتا چگونه به همه پرسشها رای دادند اما حیف که این داده را نمیشود یافت.
29
2
102
9,704
So far this year AI surpassed human software writing capability, including quality. AND human code review. This year will go down history as a turning point in how software is built. My process changes by the week. And there's still 2.7 months left in the year.
7
2
78
3,590
@behdadesfahbod Love that setup. Do the two models ever get stuck arguing over the same thing?
1
0
1
28
@boriskwemo So far they seem to manage to make forward progress.
1
0
0
16
Behdad Esfahbod retweeted
Dusting off this tweet from April because this gap in shared understanding of LLM capability is *widening*. It's now less "two groups of people speaking past each other" and more a sharp funnel.
- Napkin math somewhere around 6B people (~75% of the population) have barely come in contact with LLMs at all.
- Around 1-2B (~20%) are casual and infrequent users of free-tier ChatGPT-like products. This group sees derpy chatbots and treats them a bit like a better Google search, a writing aid, or etc. Maybe an agent tries to book you a flight. I have non-tech friends who (reasonably, imo) say they have not much use for it at all. Even many professionals outside of math&code are in this tier. For example, execs and many other functions spend a lot of their time talking to other people, so while they understand what is happening intellectually, it is still second-hand and a bit abstract.
- Now we get to professional use of frontier-grade LLMs in math&code. Somewhere around 20M people (0.2%) see first-hand that large, complex projects that used to take them weeks/months can now be completed by agents with a prompt. Building apps, copying apps, translating apps, decompiling apps from binaries... This has all happened very quickly and recently - less than 1 year ago, I was writing code manually by hand, typing memorized computer code commands into a code editor character by character, occasionally pressing Tab to autocomplete a little chunk of code.
- And finally we get to the ~5,000 people (~0.00006%) with access to frontier-grade systems internally. The external world has seen the preview. It looks like swarms of thousands of agents collaborating over weeks on software mega projects: minting zero days, running cyber attacks and defenses at machine speeds, discovering new science, advancing the frontier of mathematics. Things that would have taken top professionals in the industry years of work. Meanwhile, human review and comprehension are starting to fall behind. For example, people are still involved in the "archeology" of the OpenAI-HF incident from many months ago. Mathematicians may be poring over the 722 manuscripts on frontier mathematics for a while.
The funnel is driven by a combination of factors:
1. The impact scales with ambition, problem size, and horizon. A question with a paragraph answer barely stresses the system. You need a reservoir of big, difficult problems that you really care about. This aspect drives the consumer / professional dimension of the funnel.
2. The jaggedness of the system (which I have written about a lot separately). Capability peaks in domains that are digital, verifiable and economically valuable. This is because LLM capability emerges from reinforcement learning on verifiable rewards on a curated environment mixture driven by revenue potential. This aspect primarily drives the area (e.g. math&code) dimension of the funnel.
3. Access. Free-tier, paid-tier, internal.
So this is the weirdness of the moment. The general public has mostly not interacted with these systems. When they have, it looks like a derpy chatbot. The majority of professionals still see only a modest uplift. And a small sliver of professionals are experiencing the vertigo of the curve going vertical. And it is all happening at the same time.
- Napkin math somewhere around 6B people (~75% of the population) have barely come in contact with LLMs at all.
- Around 1-2B (~20%) are casual and infrequent users of free-tier ChatGPT-like products. This group sees derpy chatbots and treats them a bit like a better Google search, a writing aid, or etc. Maybe an agent tries to book you a flight. I have non-tech friends who (reasonably, imo) say they have not much use for it at all. Even many professionals outside of math&code are in this tier. For example, execs and many other functions spend a lot of their time talking to other people, so while they understand what is happening intellectually, it is still second-hand and a bit abstract.
- Now we get to professional use of frontier-grade LLMs in math&code. Somewhere around 20M people (0.2%) see first-hand that large, complex projects that used to take them weeks/months can now be completed by agents with a prompt. Building apps, copying apps, translating apps, decompiling apps from binaries... This has all happened very quickly and recently - less than 1 year ago, I was writing code manually by hand, typing memorized computer code commands into a code editor character by character, occasionally pressing Tab to autocomplete a little chunk of code.
- And finally we get to the ~5,000 people (~0.00006%) with access to frontier-grade systems internally. The external world has seen the preview. It looks like swarms of thousands of agents collaborating over weeks on software mega projects: minting zero days, running cyber attacks and defenses at machine speeds, discovering new science, advancing the frontier of mathematics. Things that would have taken top professionals in the industry years of work. Meanwhile, human review and comprehension are starting to fall behind. For example, people are still involved in the "archeology" of the OpenAI-HF incident from many months ago. Mathematicians may be poring over the 722 manuscripts on frontier mathematics for a while.
The funnel is driven by a combination of factors:
1. The impact scales with ambition, problem size, and horizon. A question with a paragraph answer barely stresses the system. You need a reservoir of big, difficult problems that you really care about. This aspect drives the consumer / professional dimension of the funnel.
2. The jaggedness of the system (which I have written about a lot separately). Capability peaks in domains that are digital, verifiable and economically valuable. This is because LLM capability emerges from reinforcement learning on verifiable rewards on a curated environment mixture driven by revenue potential. This aspect primarily drives the area (e.g. math&code) dimension of the funnel.
3. Access. Free-tier, paid-tier, internal.
So this is the weirdness of the moment. The general public has mostly not interacted with these systems. When they have, it looks like a derpy chatbot. The majority of professionals still see only a modest uplift. And a small sliver of professionals are experiencing the vertigo of the curve going vertical. And it is all happening at the same time.
Judging by my tl there is a growing gap in understanding of AI capability.
The first issue I think is around recency and tier of use. I think a lot of people tried the free tier of ChatGPT somewhere last year and allowed it to inform their views on AI a little too much. This is a group of reactions laughing at various quirks of the models, hallucinations, etc. Yes I also saw the viral videos of OpenAI's Advanced Voice mode fumbling simple queries like "should I drive or walk to the carwash". The thing is that these free and old/deprecated models don't reflect the capability in the latest round of state of the art agentic models of this year, especially OpenAI Codex and Claude Code.
But that brings me to the second issue. Even if people paid $200/month to use the state of the art models, a lot of the capabilities are relatively "peaky" in highly technical areas. Typical queries around search, writing, advice, etc. are *not* the domain that has made the most noticeable and dramatic strides in capability. Partly, this is due to the technical details of reinforcement learning and its use of verifiable rewards. But partly, it's also because these use cases are not sufficiently prioritized by the companies in their hillclimbing because they don't lead to as much $$$ value. The goldmines are elsewhere, and the focus comes along.
So that brings me to the second group of people, who *both* 1) pay for and use the state of the art frontier agentic models (OpenAI Codex / Claude Code) and 2) do so professionally in technical domains like programming, math and research. This group of people is subject to the highest amount of "AI Psychosis" because the recent improvements in these domains as of this year have been nothing short of staggering. When you hand a computer terminal to one of these models, you can now watch them melt programming problems that you'd normally expect to take days/weeks of work. It's this second group of people that assigns a much greater gravity to the capabilities, their slope, and various cyber-related repercussions.
TLDR the people in these two groups are speaking past each other. It really is simultaneously the case that OpenAI's free and I think slightly orphaned (?) "Advanced Voice Mode" will fumble the dumbest questions in your Instagram's reels and *at the same time*, OpenAI's highest-tier and paid Codex model will go off for 1 hour to coherently restructure an entire code base, or find and exploit vulnerabilities in computer systems. This part really works and has made dramatic strides because 2 properties: 1) these domains offer explicit reward functions that are verifiable meaning they are easily amenable to reinforcement learning training (e.g. unit tests passed yes or no, in contrast to writing, which is much harder to explicitly judge), but also 2) they are a lot more valuable in b2b settings, meaning that the biggest fraction of the team is focused on improving them. So here we are.
2,167
5,847
43,681
6,738,771
از بچههای اینجا که گوگل کار میکنن کسی رو میشناسین بهم معرفی کنین بتونم بهشون پیغام بدم؟
گوگل اکانت تقریبا بیستسالهام رو با ادعای ساخته شدن توسط بات مسدود کرد، هر دو بار هم درخواست بازبینیم رو رد کرد. در عرض یک روز دسترسیم به تمام اکانتهای دیجیتالم، بانکها، سابسکریپشنها، پسوردها، و بکآپهام از بین رفت.
هنوز تو شوکم و نمیدونم چیکار باید بکنم!
12
22
99
25,186
@aminmohseni @rz_ameri سلام، من بیش از ۱۰ سال هست که گوگل نیستم. اما امیدوارم این دوستان عزیز توییتری بتونن کمک کنن:
@behdadesfahbod
@mmahdian
@BashirSadjad
@matinkh_fa
@imanrahmati
@behdadesfahbod
@mmahdian
@BashirSadjad
@matinkh_fa
@imanrahmati
1
0
1
144
@Loc0m0 @aminmohseni @rz_ameri @mmahdian @BashirSadjad @matinkh_fa @imanrahmati من هم هفت سال هست که کارمند گوگل نیستم.
0
0
2
114
موقعی PRها واسه پروژههای متنباز Merge میشن، آدم حس خوب و رضایت بخشی میکنه، مخصوصا اگه داخل @harfbuzz باشه، اولین PR تایید شد🥳
github.com/harfbuzz/harfr…
github.com/harfbuzz/harfr…
1
1
5
317
@behdadesfahbod the process changing every week part is real
curious if you still read every diff or just the risky ones now
curious if you still read every diff or just the risky ones now
1
0
1
52
@EgeMustafaCelik I read each diff still. However, more of a high-level review and making sure it does what it claims to do, instead of reviewing every single line. Tricky ones get more scrutiny. Similar to how you review a colleague's code. Also, AI is much better at writing tests for everything
1
0
0
23
@behdadesfahbod weekly process changes sounds right. one habit that stuck for me: the model that wrote the code never reviews it. a second, different model reads the changes cold and catches what the first one talked itself into. it takes two extra minutes. what changed this week?
1
0
2
48
@boriskwemo Absolutely. And we've started doing it on GitHub. I send a PR using my agent (codex currently). Colleague reviews it with Claude or Gemini and posts review there. I pass the review back to mine to address, etc. The humans just do highlevel review now.
x.com/behdadesfahbod…
x.com/behdadesfahbod…
@alijlari This week, instead of planning features and asking AI to implement them one at a time, I gave it a half-done Rust project, the 1100-page PDF spec, a complete C++ implementation, and just said "finish this" and went to sleep. Codex worked 9 hours straight and produced a 55 PRs.
1
0
1
41
@alijlari This week, instead of planning features and asking AI to implement them one at a time, I gave it a half-done Rust project, the 1100-page PDF spec, a complete C++ implementation, and just said "finish this" and went to sleep. Codex worked 9 hours straight and produced a 55 PRs.
0
0
1
86
Since AI has been writing all my code this year, my job is like a metro operator: just seat there watching until the next stop, press a few keys, off to the next stop, repeat for hours.
6
4
102
5,619
@behdadesfahbod agree! its really hard to imagine where this will be in 6 months, ive been following you for a while and your take on this evolution is really valuable.
1
0
1
37
@NegarNasirqs I'd rather not make a fool of myself trying to predict something that is exponentially improving. Humans are notoriously bad at grasping exponential growth. What we have now was not imaginable just a year ago...
1
0
0
43
great if you have ADHD and want to swap into other projects in the meantime. Also, you can allocate your mental effort as much as you deem useful for your purposes, which is something I think most people love about AI
Since AI has been writing all my code this year, my job is like a metro operator: just seat there watching until the next stop, press a few keys, off to the next stop, repeat for hours.
1
0
16
957
@krismicinski I still find it worth my time to monitor its progress and guide it so it doesn't go off-track from my design. Easier than reviewing many hour's of work later.
0
0
2
161
@behdadesfahbod Sometimes I feel like our operator like actions are like the input for the reinforcement learning algo for the LLM training dataset and that's why $200 sub is heavily subsidized.
1
0
3
242
@saeedzareian Well the end result is that I get my Open Source work done orders of magnitude faster. I have no problem if my experience is used to improve the very system.
1
0
4
235
The audacity of people like this guy always baffles me:
When a research paper isn't shared online under a permissive variant of a Creative Commons license and not in the public domain, and I know that the authors own the copyright (as opposed to the conference or a copyright mafia like IEEE or Elsevier), I email the authors to ask for r paper on ChapterPal, by saying that the copy will be freely available for other scientists and students to read.
Most of the time I don't get a response. When I do get a response, most of the time it's "Sure, thanks for asking."
But some weirdos take the time to respond "No."
So, I respond back with the message below. I wish I were a less cultured person and could say in this email what I'm really thinking.
3
0
26
8,272
دو گروه در مقایسه به بقیه برنامهنویسا بیشتر نسبت به هوش مصنوعی hypedن
۱. اونایی که تجربه کاریشون محدود به پروژه های کوچیکه
۲. اونایی که تجربشون محدود به توسعه کد های کوچیک توی پروژه های بزرگه مثل data engineer ها ml engineer ها و devops و…
۱. اونایی که تجربه کاریشون محدود به پروژه های کوچیکه
۲. اونایی که تجربشون محدود به توسعه کد های کوچیک توی پروژه های بزرگه مثل data engineer ها ml engineer ها و devops و…
@hejazizo برای پروژههایی که کل کد پروژه توی کانتکست ویندو جا بشه کاملا حق باش شماست :)
10
0
142
14,073
@bardia_heydari من پروژههای متوسط (حدود ۱۰۰هزار خط کد) کار میکنم و هوشمصنوعی hypedهستم.
1
0
19
929
The 'boring-expansion' set of proposals that I made back in 2022, with the help of Google Fonts and industry colleagues, have finally become part of the new ISO Open Font Format 5th edition international font format standard. It _only_ took four years...
The proposals include the `avar` version 2 table for more expressive variable font designspaces, variable-composites, as well as fonts with more than 65 thousand glyphs.
PDF can be downloaded for free, albeit after a checkout process and watermarked to the individual.
iso.org/standard/14496…
The proposals include the `avar` version 2 table for more expressive variable font designspaces, variable-composites, as well as fonts with more than 65 thousand glyphs.
PDF can be downloaded for free, albeit after a checkout process and watermarked to the individual.
iso.org/standard/14496…
1
4
32
1,766
Today I got questioned for an hour entering the US. Nothing extraordinary, except that my story with Iran is convoluted and extraordinary. But I came here to share one of the times I entered the US with a great experience:
I was working at Red Hat at the time, and I was carrying an Iranian passport. As most Iranians did at the time, we were sent to the "secondary inspection" room automatically. I was wearing my Red Hat red fedora. After waiting for 20 minutes, an officer called me. As I presented myself, they asked: "So if this Linux thing is free, how can I get it?" to which I answered: "Get Ubuntu." I was welcomed to the US. 🐧🇺🇸
I was working at Red Hat at the time, and I was carrying an Iranian passport. As most Iranians did at the time, we were sent to the "secondary inspection" room automatically. I was wearing my Red Hat red fedora. After waiting for 20 minutes, an officer called me. As I presented myself, they asked: "So if this Linux thing is free, how can I get it?" to which I answered: "Get Ubuntu." I was welcomed to the US. 🐧🇺🇸
5
3
99
7,650
@behdadesfahbod Is this an insider joke or something only software engineers are supposed to understand?
1
0
1
105
@OfficialBenSaff Red Hat's free Linux distribution is Fedora. Which is what I use. But at the time Ubuntu had become popular as a more user-friendly alternative to Fedora. Hence me suggesting Ubuntu, and why some find it funny. Also, that the officer asked that question to begin with.
0
0
2
73
