LESSWRONG
LW

HomeAll PostsConceptsLibrary
Best of LessWrong
Sequence Highlights
Rationality: A-Z
The Codex
HPMOR
Community Events
LW the Album
About
FAQ
HomeAll PostsConceptsLibraryCommunity
Customize
Rationality+Rationality+World Modeling+World Modeling+AIAIWorld OptimizationWorld OptimizationPracticalPracticalCommunityCommunity
Personal Blog+
404Welcome to LessWrong!
Ruby, Raemon, RobertM, habryka
48

Quick Takes

RobertM120
1
Vaguely feeling like OpenAI might be moving away from GPT-N+1 release model, for some combination of "political/frog-boiling" reasons and "scaling actually hitting a wall" reasons.  Seems relevant to note, since in the worlds where they hadn't been drip-feeding people incremental releases of slight improvements over the original GPT-4 capabilities, and instead just dropped GPT-5 (and it was as much of an improvement over 4 as 4 was over 3, or close), that might have prompted people to do an explicit orientation step.  As it is, I expect less of that kind of orientation to happen.  (Though maybe I'm speaking too soon and they will drop GPT-5 on us at some point, and it'll still manage to be a step-function improvement over whatever the latest GPT-4* model is at that point.)
Alexander Gietelink Oldenziel4-2
0
Wildlife Welfare Will Win The long arc of history bend towards gentleness and compassion. Future generations will look with horror on factory farming. And already young people are following this moral thread to its logical conclusion; turning their eyes in disgust to mother nature, red in tooth and claw. Wildlife Welfare Done Right, compassion towards our pets followed to its forceful conclusion would entail the forced uploading of all higher animals, and judging by the memetic virulences of shrimp welfare to lower animals as well.  Morality-upon-reflexion may very well converge on a simple form of pain-pleasure utilitarianism.  There are few caveats: future society is not dominated, controlled and designed by a singleton AI-supervised state, technology inevitable stalls and that invisible hand performs its inexorable logic for the eons and an Malthuso-Hansonian world will emerge once again - the industrial revolution but a short blip of cornucopia.  Perhaps a theory of consciousness is discovered and proves once and for all homo sapiens and only homo sapiens are conscious ( to a significant degree). Perhaps society will wirehead itself into blissful oblivion.  Or perhaps a superior machine intelligence arises, one whose final telos is the whole of and nothing but office supplies. Or perhaps stranger things still happen and the astronomo-cosmic compute of our cosmic endowment is engaged for mysterious purposes. Arise, self-made god of pancosmos. Thy name is UDASSA. 
Linch303
3
(x-posted from the EA Forum) We should expect that the incentives and culture for AI-focused companies to make them uniquely terrible for producing safe AGI.  From a “safety from catastrophic risk” perspective, I suspect an “AI-focused company” (e.g. Anthropic, OpenAI, Mistral) is abstractly pretty close to the worst possible organizational structure for getting us towards AGI. I have two distinct but related reasons: 1. Incentives 2. Culture From an incentives perspective, consider realistic alternative organizational structures to “AI-focused company” that nonetheless has enough firepower to host multibillion-dollar scientific/engineering projects: 1. As part of an intergovernmental effort (e.g. CERN’s Large Hadron Collider, the ISS) 2. As part of a governmental effort of a single country (e.g. Apollo Program, Manhattan Project, China’s Tiangong) 3. As part of a larger company (e.g. Google DeepMind, Meta AI) In each of those cases, I claim that there are stronger (though still not ideal) organizational incentives to slow down, pause/stop, or roll back deployment if there is sufficient evidence or reason to believe that further development can result in major catastrophe. In contrast, an AI-focused company has every incentive to go ahead on AI when the cause for pausing is uncertain, and minimal incentive to stop or even take things slowly.  From a culture perspective, I claim that without knowing any details of the specific companies, you should expect AI-focused companies to be more likely than plausible contenders to have the following cultural elements: 1. Ideological AGI Vision AI-focused companies may have a large contingent of “true believers” who are ideologically motivated to make AGI at all costs and 2. No Pre-existing Safety Culture AI-focused companies may have minimal or no strong “safety” culture where people deeply understand, have experience in, and are motivated by a desire to avoid catastrophic outcomes.  The first one should be self-explanatory. The second one is a bit more complicated, but basically I think it’s hard to have a safety-focused culture just by “wanting it” hard enough in the abstract, or by talking a big game. Instead, institutions (relatively) have more of a safe & robust culture if they have previously suffered the (large) costs of not focusing enough on safety. For example, engineers who aren’t software engineers understand fairly deep down that their mistakes can kill people, and that their predecessors’ fuck-up have indeed killed people (think bridges collapsing, airplanes falling, medicines not working, etc). Software engineers rarely have such experience. Similarly, governmental institutions have institutional memories with the problems of major historical fuckups, in a way that new startups very much don’t.
Johannes C. Mayer40
0
What are you Doing? What did you Plan? [Suno] What are you doing? What did you plan? Are they aligned? If not then comprehend, if what you are doing now is better than the original thing. Be open-minded about, what is the optimal thing. Don't fix the bottom line too: "Whatever the initial plan was is the best thing to do." There are sub-agents in your mind. You don't want to fight, with them, as usually they win in the end. You might then just feel bad and don't even understand why. As a protective skin your sub-agent hides, the reasons for why, you feel so bad right now. At that point, you need to pack the double crux out. But ideally, we want to avoid, any conflict that might arise. So don't ask yourself if you followed your consequentialist reasoner's plan. Instead just ask: "What is the best thing for me to do right now?" while taking all the sub-agents into account. To do it set a timer for 1 minute, and spend that time reflecting about: What do you want to get out of this session of work, why is this good, how does this help? You can wirte notes in advance, then document your plans, and then read them out loud. to remember the computations your brain did before, such that you don't need to repeat some of these chores. Ideally, the notes would talk about, the reasons for why something seemed like a good thing to try. But then as you evaluate what next step you could take, drop that bottom line. Treat it as evidence for what your brain computed in the past as an optimal policy, but nothing more. It's now your new goal to figure out again for yourself, using all the subagents within your shell. And to do this regularly you of course use a timer you see. Every 30 minutes to an hour it should ring out loud reminding you to evaluate, what would be the next step to take. If you let everybody influence the decision process that will commence, the probability is high that after you decide there will be no fight, in your mind.
Alexander Gietelink Oldenziel20
0
My timelines are lengthening.  I've long been a skeptic of scaling LLMs to AGI *. To me I fundamentally don't understand how this is even possible. It must be said that very smart people give this view credence. davidad, dmurfet. on the other side are vanessa kosoy and steven byrnes. When pushed proponents don't actually defend the position that a large enough transformer will create nanotech or even obsolete their job. They usually mumble something about scaffolding. I won't get into this debate here but I do want to note that my timelines have lengthened, primarily because some of the never-clearly-stated but heavily implied AI developments by proponents of very short timelines have not materialized. To be clear, it has only been a year since gpt-4 is released, and gpt-5 is around the corner. Still my timelines are lengthening.  A year ago, when gpt-3 came out progress was blindingly fast. Part of short timelines came from a sense of 'if we got surprised so hard by gpt2-3, we are completely uncalibrated, who knows what comes next'. People seemed surprised by gpt-4 in a way that seemed uncalibrated to me. gpt-4 performance was basically in line with what one would expect if the scaling laws continued to hold. At the time it was already clear that the only really important driver was compute  data and that we would run out of both shortly after gpt-4. Scaling proponents suggested this was only the beginning, that there was a whole host of innovation that would be coming. Whispers of mesa-optimizers and simulators.  One year in: Chain-of-thought doesn't actually improve things that much. External memory and super context lengths ditto. A whole list of proposed architectures seem to serve solely as a paper mill. Every month there is new hype about the latest LLM or image model. Yet they never deviate from expectations based on simple extrapolation of the scaling laws. There is only one thing that really seems to matter and that is compute and data. We have about 3 more OOMs of compute to go. Data may be milked another OOM.  A big question will be whether gpt-5 will suddenly make agentGPT work ( and to what degree). It would seem that gpt-4 is in many ways far more capable than (most or all) humans yet agentGPT is curiously bad.  All-in-all AI progress** is developing according to the naive extrapolations of Scaling Laws but nothing beyond that. The breathless twitter hype about new models is still there but it seems to be believed more at a simulacra level higher than I can parse.  Does this mean we'll hit an AI winter? No. In my model there may be only one remaining roadblock to ASI (and I suspect I know what it is). That innovation could come at anytime. I don't know how hard it is, but I suspect it is not too hard.  * the term AGI seems to denote vastly different things to different people in a way I find deeply confusing. I notice that the thing that I thought everybody meant by AGI is now being called ASI. So when I write AGI, feel free to substitute ASI.  ** or better, AI congress addendum:  since I've been quoted in dmurfet's AXRP interview as believing that there are certain kinds of reasoning that cannot be represented by transformers/LLMs I want to be clear that this is not really an accurate portrayal of my beliefs. e.g. I don't think transformers don't truly understand, are just a stochastic parrot, or in other ways can't engage in the abstract reasoning that humans do. I think this is clearly false, as seen by interacting with any frontier model. 

Popular Comments

Load More

Recent Discussion

RobertM's Shortform
RobertM
12RobertM
Vaguely feeling like OpenAI might be moving away from GPT-N+1 release model, for some combination of "political/frog-boiling" reasons and "scaling actually hitting a wall" reasons.  Seems relevant to note, since in the worlds where they hadn't been drip-feeding people incremental releases of slight improvements over the original GPT-4 capabilities, and instead just dropped GPT-5 (and it was as much of an improvement over 4 as 4 was over 3, or close), that might have prompted people to do an explicit orientation step.  As it is, I expect less of that kind of orientation to happen.  (Though maybe I'm speaking too soon and they will drop GPT-5 on us at some point, and it'll still manage to be a step-function improvement over whatever the latest GPT-4* model is at that point.)
Raemon20

Yeah. This prompts me to make a brief version of a post I'd had on my TODO list for awhile:

"In the 21st century, being quick and competent at 'orienting' is one of the most important skills." 

(in the OODA Loop sense, i.e. observe -> orient -> decide -> act)

We don't know exactly what's coming with AI or other technologies, we can make plans informed by our best-guesses, but we should be on the lookout for things that should prompt some kind of strategic orientation. @jacobjacob has helped prioritize noticing things like "LLMs are pretty soon g... (read more)

Reply
Benefitial habits/personal rules with very minimal tradeoffs?
15
Slapstick

I'm looking for personal rules one might live by which adhere to a specific criteria outlined below, following an example.

I have a personal rule I've been following which is "No looking at screens in the bed where I sleep" I find this to be an extremely helpful and successful rule despite being someone who struggles to impose rules on myself.

I think one main reason it's successful for me is there aren't really any meaningful tradeoffs I'm making. If I'm feeling a need/compulsion to comfortably self soothe on my phone I can use another piece of furniture.

This rule doesn't ever result in bargaining with myself until I concede to breaking the rule, even at my lowest I'm easily appeased, just so long as it's in a slightly different...

(See More – 164 more words)
Answer by Brendan Long20

I'm not sure if this is quite what you're looking for, but one thing I do is store things I need for work and travel in consistent bags.

For example, my work laptop and my badge live in specific parts of a work-specific backpack, and I never the leave the badge anywhere except in the specific pocket of the backpack or attached to my belt.

For travel, I keep a couple things that are annoying to forget in my carry-on bag (travel-sized soap, conditioner, toothpaste, an un-opened toothbrush, a multi-country power-adapter and a spare swim suit). A battery and ant... (read more)

Reply
2Slapstick
That one sounds good! It wouldn't work for me personally because I have a pathological relationship with refined sugar so the only equilibrium which works for me is cutting it out entirely (which has been successful and rewarding though initially very difficult). Thanks!
2ErioirE
I'm in a similar situation. I have very little self control with sweets/candy if I have them available. I can far more easily stop myself from buying them in the first place. If I allow myself to buy a bag of candy I've already lost and I will consume all of it in a matter of hours/days.
1Slapstick
Oh that's a good one! I mostly follow that one already although I do find value in some unsweetened teas and smoothies. I find personally that the immediate trade-offs to consuming alcohol are enough to ensure I only really drink when it's actually aligned with my interests. Although I do have a rule for alcohol which is "don't consume any alcohol unless people who you're currently being social with are already drinking," I'm not sure exactly how much that rule has helped me because I've followed it all my life and I don't really like alcohol that much, but maybe that's partially because of the rule. But yes I think the rule you gave is a really good one, especially when it comes to things like refined sugar. A sugar craving could be satisfied in other ways, so there's relatively small trade-offs in that sense, whereas it's very beneficial not to drink refined calories because it's so easy to consume so much that way while not bringing in any significant nutrition alongside it. Thanks!
What comes after Roam's renaissance?
12
Itay Dreyfus
This is a linkpost for https://productidentity.co/p/4-roam-research-what-comes-after

A new year has come. It's 2024 and note-taking isn’t cool anymore. The once-blooming space has had its moment. Moreover, the almighty Roam Research isn’t the only king anymore.

The hype is officially over.

At this time of year, when many are busy reflecting on the past year while excitingly looking into the future, I realized it's a good opportunity to look back at Roam’s madness timeline. The company that took Twitterverse and Silicon Valley by storm is now long after its breakthrough.

Roam was one of those phenomena that happen every other few years. Its appearance in our lives not only made the “tools for thought” niche fashionable. It marked a new era in the land of note-taking apps. In conjunction with a flourishing movement of internet intellectuals[1], it...

(Continue Reading – 3138 more words)
Vlad Sitalo10

I’m using a different but similar approach of incorporating SRS into Roam instead ( https://vlad.roam.garden/apply-Spaced-Repetition-to-evergreen-notes-that-you-want-to-remember-or-periodically-rise-to-attention  ). 

Reply
1Itay Dreyfus
From my personal perspective, I think Roam brought to attention a different type of software and vibe which were dominated SV type startups.  Through Roam, I've learned about all those fancy-old ideas and concepts. I guess I was coming from a more mainstream corner of the tech scene. My excitement declined as the hype too, and as you describe, I couldn't understand its real benefits as I didn't have a strong enough reason to use it.  Coming to Roam a few years later, after starting a publication, has made the difference as I'm less focused on [[double-bracketing]] but just taking notes, gradually and only when I'm feeling like it.
Can we build a better Public Doublecrux?
52
Raemon

Something I'd like to try at LessOnline is to somehow iterate on the "Public Doublecrux" format. I'm not sure if I'll end up focusing on it, but here are some ideas.

Public Doublecrux is a more truthseeking oriented version of Public Debate. The goal of a debate is to change your opponent's mind or the public's mind. The goal of a doublecrux is more like "work with your partner to figure out if you should change your mind, and vice versa."

Reasons to want to do public doublecrux include:

  • It helps showcase subtle mental moves that are hard to write down explicitly (i.e. tacit knowledge transfer.
  • There's still something good and exciting about seeing high profile smart people talk about ideas. Having some variant of that format seems good for LessOnline.
...
(Continue Reading – 1179 more words)
DPiepgrass10

Doublecrux sounds like a better thing than debate, but why such an event should be live? (apart from "it saves money/time not to postprocess")

Reply
1scarcegreengrass
I quite like the Arguman format of flowcharts to depict topics. In a live performance, participants might sometimes add nodes to the flowchart, or sometimes ask for revision to another participant's existing node. For example, asking for rewording for clarity. Perhaps the better term would be tree, not flowchart. Each node is a response to its parent. This could perhaps be implemented with bulleted lists in a Google Doc. It's nice for the event to output a useful document.
2whestler
I think it might be a good idea to classify a "successful" double crux as being a double crux where both participants agree on the truth of the matter at the end, or at least have shifted their world views to be significantly more coherent. It seems like the main obstacles to successful double crux are emotional (pride, embarrassment), and associations with debates, which threaten to turn the format into a dominance contest. It might help to start with a public and joint announcement by both participants that they intend to work together to discover the truth, recognising that their currently differing models means that at least one of them has the opportunity to grow in their understanding of the world and become a stronger rationalist, that they are committed to helping each other become stronger in the art. Alternatively you could have the participants do the double crux in their own time, and in private (though recorded). If the double crux succeeds, then post it, and major kudos to the participants. If it fails, then simply post the fact that the crux failed but don't post the content. If this format is used regularly, eventually it may become clear which participants consistently succeed in their double crux attempts, and which don't, and they can build reputation that way, rather than trying to "win" a debate.
Alexander Gietelink Oldenziel's Shortform
Alexander Gietelink Oldenziel
Alexander Gietelink Oldenziel20

My timelines are lengthening. 

I've long been a skeptic of scaling LLMs to AGI *. To me I fundamentally don't understand how this is even possible. It must be said that very smart people give this view credence. davidad, dmurfet. on the other side are vanessa kosoy and steven byrnes. When pushed proponents don't actually defend the position that a large enough transformer will create nanotech or even obsolete their job. They usually mumble something about scaffolding.

I won't get into this debate here but I do want to note that my timelines have lengthe... (read more)

Reply
4Alexander Gietelink Oldenziel
Wildlife Welfare Will Win The long arc of history bend towards gentleness and compassion. Future generations will look with horror on factory farming. And already young people are following this moral thread to its logical conclusion; turning their eyes in disgust to mother nature, red in tooth and claw. Wildlife Welfare Done Right, compassion towards our pets followed to its forceful conclusion would entail the forced uploading of all higher animals, and judging by the memetic virulences of shrimp welfare to lower animals as well.  Morality-upon-reflexion may very well converge on a simple form of pain-pleasure utilitarianism.  There are few caveats: future society is not dominated, controlled and designed by a singleton AI-supervised state, technology inevitable stalls and that invisible hand performs its inexorable logic for the eons and an Malthuso-Hansonian world will emerge once again - the industrial revolution but a short blip of cornucopia.  Perhaps a theory of consciousness is discovered and proves once and for all homo sapiens and only homo sapiens are conscious ( to a significant degree). Perhaps society will wirehead itself into blissful oblivion.  Or perhaps a superior machine intelligence arises, one whose final telos is the whole of and nothing but office supplies. Or perhaps stranger things still happen and the astronomo-cosmic compute of our cosmic endowment is engaged for mysterious purposes. Arise, self-made god of pancosmos. Thy name is UDASSA. 
Questions are usually too cheap
46
Nathan Young
This is a linkpost for https://nathanpmyoung.substack.com/p/questions-are-too-cheap

It is easier to ask than to answer. 

That’s my whole point.

It is much cheaper to ask questions than answer them so beware of situations where it is implied that asking and answering are equal. 

Here are some examples:

Let's say there is a maths game. I get a minute to ask questions. You get a minute to answer them. If you answer them all correctly, you win, if not, I do. Who will win?

Preregister your answer.

Okay, let's try. These questions took me roughly a minute to come up with. 

What's 56,789 * 45,387?

What's the integral from -6 to 5π of sin(x cos^2(x))/tan(x^9) dx?

What's the prime factorisation of 91435293173907507525437560876902107167279548147799415693153?

Good luck. If I understand correctly, that last one's gonna take you at least an hour1 (or however long it takes to threaten...

(Continue Reading – 1669 more words)
Nathan Young20

Though sometimes the obligation to answer is right, right? I guess maybe it's that obligation works well at some scale, but then becomes bad at some larger scale. In a coversation, it's fine, in a public debate, sometimes it seems to me that it doesn't work.

Reply
2Nathan Young
I think the motivating instances are largely: * Online debates are bad * Freedom Of Information requests suck I think I probably backfilled from there. I do sometimes get persistant questions on twitter, but I don't think there is a single strong example.
1Pavgran
109647247078573083699910710287 × 833904139046784164224502687119 With the right tool, it takes about 12 seconds. 5 to locate the tool, 7 for it to give the answer. Just nitpicking, of course. You could have easily taken 60-digit primes.
2Nathan Young
Sadly you are the second person to correct me on this @Paul Crowley was first. Ooops. 
To get the best posts emailed to you, create an account! (2-3 posts per week, selected by the LessWrong moderation team.)
Log In Reset Password
...or continue with
GOOGLEGITHUB
Against Student Debt Cancellation From All Sides of the Political Compass
6
Maxwell Tabarrok
This is a linkpost for https://www.maximum-progress.com/p/against-student-debt-cancellation

A stance against student debt cancellation doesn’t rely on the assumptions of any single ideology. Strong cases against student debt cancellation can be made based on the fundamental values of any section of the political compass. In no particular order, here are some arguments against student debt cancellation from the perspectives of many disparate ideologies.

Equity and Fairness

Student debt cancellation is a massive subsidy to an already prosperous and privileged population. American college graduates have nearly double the income of high school graduates. African Americans are far underrepresented among degree holders compared to their overall population share.

Within the group of college graduates debt cancellation increases equity, but you can’t get around the fact that 72% of African Americans have no student debt because they never went to college....

(Continue Reading – 1331 more words)
Algon20

Wow, I didn't realize just how bad student debt cancellation is from so many perspectives. Now I want more policy critiques like this. 

Reply
What you really mean when you claim to support “UBI for job automation”
17
Deric Cheng

Author’s Note: Though I’m currently a governance researcher at Convergence Analysis, this post is unaffiliated with Convergence. The opinions expressed are solely my own.

You’ve seen it a dozen times at this point. You’re probably broadly aligned philosophically, but haven’t thought terribly deeply about the details. You generally support Andrew Yang’s $12k / year “Freedom Dividend” as “moving in the right direction”, even if it’s economically flawed.

The argument goes roughly like this: “All of our jobs are about to be automated away with AI technology and robotics! We’ll end up soon in a post-work society with massive unemployment unless we can find a way to distribute the benefits of AI automation fairly. We need a universal basic income to protect humans.”

To recap - universal basic income is a...

(Continue Reading – 2789 more words)
FlorianH10

Commenting on the basis of lessons from some experience doing UBI analysis for Switzerland/Europe:

The current systems has various costs (time and money, but maybe more importantly, opportunities wasted by perverse incentives) associated with proving that you are eligible for some benefit.

On the one hand, yes, and its a key reason why NIT/UBI systems are often popular on the right; even Milton Friedman already advocated for a NIT. That said, there are also discussions that suggest the poverty trap - i.e. overwhelmingly strong labor disincentives for poor, f... (read more)

Reply
1ErioirE
As a software developer who works on object-level automation every day, I'm intimidated by the difficulty of attempting to definitively quantify 'profit from automated tasks' in a useful way. For example, how do we define 'automation'? "A task that formerly needed to be done by a human that now doesn't need to be"? A printing press is automation by some interpretations of that insufficient definition. Some changes in efficiency also have similar effects on productivity without being 'automation' (although much less scalable), for example a user that becomes highly proficient in the hotkeys of a complex platform may see massive improvements in their productivity, and subsequently eliminate jobs that would have been needed if they hadn't become more productive. I suspect if additional taxes were levied on 'job automation' it would merely create large incentives for companies to skirt around whatever the legal definition of automation was, and potentially hide it in things like the above example. In the case where there was no 'automation tax' created, I would anticipate a NIT to be reasonable but not sustainable because I expect automation to continue to remove jobs at an accelerating rate in years to come. I do not expect tax revenue to increase at the same rate because my current understanding is that the most wealthy tend to also be those most proficient at exploiting loopholes in the tax system to evade as much as possible. My takes here are almost entirely conjecture and I'd appreciate someone more informed to correct and/or clarify.
2Seth Herd
Thanks for doing this! It's looking like we may need major economic changes to keep up with job automation (assuming we don't get an outright AGI takeover). So, getting started on thinking this stuff through may have immense benefit. Like the alignment problem, it's embarassing as a species that we haven't thought about this more when the train appears to be barreling down the tracks. So, kudos and keep it up! Now, the critique: doing this analysis for only the richest country in the world seems obviously inadequate and not even a good starting point; something like the median country would be more useful. OTOH, I see why you're doing this; I'm a US citizen and numbers are easier to get here. So in sum, I think the bigger issue is the second one you mention: global tax reform that can actually capture the profits made from various AI companies and the much larger base of AI-enabled companies that don't pay nearly as much for AI as they would for labor, but reap massive profits. They will often be "based" in whatever country gives them the lowest tax rates. So we have another thorny global coordination problem. I was also going to mention not accounting for the tech changes this is accounting for. So I recommend you add that this is part 1 in the intro to head off that frustration among readers.
1Deric Cheng
Totally agree on the UBI being equivalent to a negative income tax in many ways! My main argument here is that UBI is a non-realistic policy when you actually practically implementing it, whereas NIT is the same general outcome but significantly more realistic. If you use the phrase UBI as the "high-level vision" and actually mean "implement it as a NIT" in terms of policy, I can get behind that. Re: the simplicity idea, repeating what I left in a comment above:  Personally, I really don't get the "easy to maintain" argument for UBI, esp. given my analysis above. You'd rather have a program that costs $4 trillion with zero maintenance costs, than a similarly impactful program that costs $~650 billion with maintenance costs? It's kind of a reductive argument that only makes sense when you don't look at the actual numbers behind implementing a policy idea.
OpenAI releases GPT-4o, natively interfacing with text, voice and vision
19
Martín Soto
This is a linkpost for https://openai.com/index/gpt-4o-and-more-tools-to-chatgpt-free/
  • Until now ChatGPT dealt with audio through a pipeline of 3 models: audio transcription, then GPT-4, then text-to-speech. GPT-4o is apparently trained on text, voice and vision so that everything is done natively. You can now interrupt it mid-sentence.
  • It has GPT-4 level intelligence according to benchmarks. Somewhat better at transcription than Whisper, and considerably better at vision than previous models.
  • It's also somehow been made significantly faster at inference time. Might be mainly driven by an improved tokenizer. Edit: Nope, English tokenizer is only 1.1x.
  • It's confirmed it was the "gpt2" model found at LMSys arena these past weeks, a marketing move. It has the highest ELO as of now.
  • They'll be gradually releasing it for everyone, even free users.
  • Safety-wise, they claim to have run it through their Preparedness framework
...
(See More – 35 more words)
Neel Nanda70

Might be mainly driven by an improved tokenizer.

I would be shocked if this is the main driver, they claim that English only has 1.1x fewer tokens, but seem to claim much bigger speed-ups

Reply
11Zach Stein-Perlman
I'm disappointed and I think they shouldn't get much credit PF-wise: they haven't published their evals, published a report on results, or even published a high-level "scorecard." They are not yet meeting the commitments in their beta Preparedness Framework — some stuff is unclear but at the least publishing the scorecard is a commitment. (Also: it's now been six months since they published the beta Preparedness Framework!)

LessOnline Festival

Ticket prices increase in 1 day

Join us May 31st to June 2nd, at Lighthaven, Berkeley CA