Rendered at 09:01:27 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
simonw 11 hours ago [-]
> [...] Let's maybe add a helmet? It could improve riding theme, but may obscure head. Maybe a small cycling cap or helmet? The user didn't ask; can add red helmet? Might be cute. But pelican with big beak; a helmet might obscure. Better maybe no.
No one’s talking about how good the final product is.
Edit: someone else commented that as I was typing this, lol.
blackhaz 2 hours ago [-]
I wonder, do we need a new benchmark? There's quite a bit of feedback data floating around about pelicans on bicycles already.
gs17 9 hours ago [-]
> Let's maybe add comments? The final code can have comments. Fine.
chvid 3 hours ago [-]
This is a remarkable coherent and clear reasoning trace.
Maybe you should start also comparing reasoning traces when you do your pelican benchmark.
armcat 2 hours ago [-]
That would be very interesting but only the open models allow you to see the reasoning trace
tyre 8 hours ago [-]
This is actually pretty good!
kurante 10 hours ago [-]
Is the broken English an optimization or a byproduct of the model being developed in China?
acheong08 10 hours ago [-]
When GPT-5.6-sol's reasoning traces were leaked, they also used "caveman speak". Definitely a token efficiency optimization
beefsack 10 hours ago [-]
I can't help but imagine agents using caveman speak sometimes start behaving in a stereotypically caveman manner, even if it's subtle. Is there a chance the agent does less reasoning because of it?
dotancohen 2 hours ago [-]
Caveman invented fire, the wheel, domesticated wild plants and animals, organised society, survived the Toba catastrophe, cooked food, and was having sex ages before you and me. Don't write him off as stupid.
It's not exactly a joke, it does reduce the amount of tokens. However, it does not improve performance (fine tunes are finnecky things, hard to get one right).
walrus01 8 hours ago [-]
Personally the only 'enthusiast' modified qwen 3.6 27b or 3.6 35b-a3b I've found useful are the ones that have been run through heretic and adversarial data sets for innocent/dangerous prompts, to produce uncensored LLMs. They have some niche non-coding uses for things that a commercial LLM will never talk about.
I think those are mostly vapor that runs on the small culture of "models should not be censored" thing. But from my experience, they unlock nothing meaningful.
Fine-tuning is great for really small models on specific applications, but it's not something that can essentially improve a more generic model.
That said, there seems to be a fine line in quantization+finetuning that could recover performance. It's just hard to get a hold of it (I feel it in some models, but it's hard to say yet; lots of small labs working on this RN).
walrus01 8 hours ago [-]
The most interesting use I've found for them so far is strictly as a novelty. Give a chat session with one to a completely non technical person, who at least knows that openai and anthropic have some guard rails on stuff, and tell them to wild with something like "give me the precursors and chemical formulas for the precusors for crystal meth" and watch it answer.
dotancohen 2 hours ago [-]
But does it answer those queries correctly, or does it just not refuse to not halucinate an incorrect answer? From where would it even have that information?
gaigalas 4 hours ago [-]
Yep, but that's not changing the quality of the model. It's not an optimization in any sense (and it's a hit on productive workflows, possibly).
This is also likely to stop working as censoring moves to the training data source.
fc417fc802 8 hours ago [-]
Training a variant to reason in early modern english in the style of the tudor elites might be an amusing way to test for that.
Barbing 9 hours ago [-]
"Neuralese"
altmanaltman 6 hours ago [-]
Just so we are clear, no "caveman" spoke English. "Caveman speak" is just shortening the vocabulary of english, not a "caveman language". Given this, your concerns for "stereotypical caveman manner" makes very little sense since what caveman are you talking about?
dotancohen 2 hours ago [-]
The concern is not that the model was trained on actual caveman artifacts, rather on modern media representations of the stereotypical caveman (that never actually existed).
armcat 17 minutes ago [-]
Less tokens. These models already overthink like crazy especially for complex tasks.
minimaxir 10 hours ago [-]
Optimization. Why use many word when few word do trick?
andsoitis 9 hours ago [-]
> Why use many word when few word do trick?
Be concise.
OR
Brief is best.
OR
Eschew verbosity
etc.
gjvc 8 hours ago [-]
"Omit needless words."
-- William Strunk Jr. and E.B. White., The Elements of Style
TiredOfLife 58 minutes ago [-]
See world
stavros 10 hours ago [-]
What I find funny about "why use many word when few word do trick?" is that it's only slightly shorter than the regular "why use many words when few words do the trick?"
inopinatus 1 hours ago [-]
The latter is not a complete alternative, it is ambiguously conflating vocabulary scale with word count, and also, it is not as funny
rapind 9 hours ago [-]
I always figured that was part of the joke, because a writer came up with it, and a writer would know (I assume?).
gaigalas 8 hours ago [-]
Optimization on a idiosyncrasy. The same thing that makes Claude repeat "That was the most important thing you said in this whole conversation" is what makes grug speak optimize on token usage.
Real humans get non-primary information from word variation. It's reasonable to hypothesize that it has a role in thinking things, because it endures. Our languages need to breathe over time, and flourishing might be one of the aspects that allows that breathing space.
walrus01 9 hours ago [-]
qwen3.8-flash-next also 'thinks' like this in its thinking stage before output, watching it 'think' in opencode, but it produces syntax correct and grammatically correct code comments, changelogs and readme type files.
ekianjo 10 hours ago [-]
Saving tokens
andsoitis 9 hours ago [-]
More intelligent and shorter:
Maybe add a small cycling cap or helmet if it doesn’t obscure the head.
AdamConwayIE 7 hours ago [-]
Likely something that was first made especially obvious by Chinese models and then became something worth optimizing for in English too.
Chinese can be extremely information-dense in token terms, though it depends on the tokenizer. Roughly speaking, you can pack more "meaning" into a short sequence than English often allows for. That's why "caveman" reasoning is a pretty good fit.
There's a difference between bolting caveman speak onto an existing model and training a model to reason that way, though. If you just force an existing model to be concise in outputs, you're artificially reducing its available reasoning steps and can possibly prevent useful exploration or verification. If it's trained specifically to use compressed reasoning, it can learn to represent the same intermediate ideas in fewer generated tokens, cutting the number of sequential inference steps without necessarily sacrificing the useful reasoning itself.
It's not so much inherently a Chinese-model trait, but Chinese models could definitely have helped demonstrate how effective very compressed reasoning traces can be.
I wouldn't say it was Chinese specifically that was emulated, but it got people thinking about tokenizers and representation efficiency, and how natural English is rather inefficient.
moezd 4 hours ago [-]
Did he seriously automate away one of the best quirks of his blog posts, i.e. evaluating new models with a touch of fun? I read AI slop all day, thanks.
KeplerBoy 45 minutes ago [-]
That's just part of the output along the SVG/image.
delichon 8 hours ago [-]
If someone can look at that reasoning trace and see a stochastic parrot next word prediction machine, we don't understand those words in the same way.
slopinthebag 50 minutes ago [-]
Why can't a next token prediction machine not predict a train of reasoning?
0xfaded 6 hours ago [-]
I still call them stochastic parrots, but believe what they are revealing is that we are all stochastic parrots to some extent. I simply don't see how biological computation (i.e. thinking) can be anything else. Similar to the reveal in west world, we are likely much simpler than we give ourselves credit for.
A "train of thought" can be seen as a trace of a depth first search where the preceding trace is used to guide termination and next expansion decisions. A similar concept, "taboo search", exists in classical constraint optimization where previous solutions are fit to a model that guides future expansion (but as the name "taboo" implies, away from uninteresting solutions).
We also have harnesses that perform breath first search.
If I tried to describe what it means to "think deeply", I would probably say a combination of both.
Ultimately I believe that we will surpass human capabilities but fail with alignment. Handing the world's resources over to stochastic systems that can evolve faster than we can reason about them simply leaves too many "interesting" outcomes that do not end well. I also expect the failure modes will be totally non-obvious.
BoredomIsFun 2 hours ago [-]
> we are all stochastic parrots to some extent.
I think this statement is continuation of the old fallacy - every generation thinks of brain in terms of what is the current technology zaitgeist is - was it 19th century when they thought brain is a network of pneumatic pipes?
vasco 3 hours ago [-]
As long as there's enough of them with different goals it doesn't matter, they'll keep each other in check. The worlds resources are already handed over to the worst people and we're still doing fine and none of the billionaires are "aligned with society". They just align with their own belly but because they want different things it all kinda works.
sujzhsbnwjek 2 hours ago [-]
These “worst people” need you. They physically need you alive to perform labor for them and to give them money (and status).
That’s the reason we are “doing fine”. Once they stop needing you..
Also, both our comments brush over the generational struggles for fairness over the centuries. We have fought to be “fine”, it did not just happen. Without fairness being introduced by force you and I would be slaving away in some sweatshop getting paid nickels as was the norm not so long ago.
Edit: That’s also assuming you are Caucasian. If you are of a different ethnicity.. well, historically, all bets are off. You could also be the literal possession of some of these “worst people” with not even your own children considered yours.
vasco 2 hours ago [-]
My only point is in a many agent system with different goals it "doesn't matter" that some agents have bad goals as long as there's enough variability of goals and resources that they can't put their vision in place.
pasteleft 4 hours ago [-]
LLM is "stochastic parrot next word prediction machine"; it's just that this "stochastic parrot next word prediction machine" have proven to be smarter than most people. I mean, this already happened with AlphaGo too.
BoredomIsFun 2 hours ago [-]
> to be smarter than most people.
Hell no. In very narrow tasks - yes, in vast majority, esp. involving state tracking (board games) and spatial reasoning - they are awful.
sneak 8 hours ago [-]
Yeah, it has been clear for a long time that there is reasoning and mental modeling going on here.
The other option is that you do understand those words the same way, and the people making these (now nonsensical) anti-AI claims simply aren’t talking about the same programs/models we are. Their idea of SOTA is when chatgpt.com launched.
If you took a point sample pre-Opus, and didn’t write a good prompt, of course you would think all AI programming was worthless slop.
negura 2 hours ago [-]
Maybe get yourself checked for chatbot psychosis. I am using current models productively, every day, and have 0 (and I mean precisely, literally 0) issue with calling it a stochastic parrot, one which lacks any kind of mentality whatsoever. There is not a shred of doubt in my mind that this is purely a statistical model, generating sequences of words, that happen to make sense in our actual mentality.
7 hours ago [-]
dnautics 7 hours ago [-]
The stochastic parrot epithet is so 4 months ago
kakadu 11 hours ago [-]
[flagged]
vessenes 10 hours ago [-]
Simon’s karma exceeds yours by about 1,300x — so most of us disagree with you. Like me, I disagree with you. I like the pelican benchmarking, and I like it when Simon stays on top of models for both public good reasons and because it saves me time and effort.
phyrex 11 hours ago [-]
Nah, keep it up. I just come to the reviews for the pelican!
andai 10 hours ago [-]
I think the pelican thing is a bit tired at this point, however I find Simon's comments to be high quality.
metmac 10 hours ago [-]
Same.
dannyw 10 hours ago [-]
Comments that the HN community find interesting are surfaced higher.
Just tap on the [-], and upvote what you find more interesting :)
jonplackett 10 hours ago [-]
If I had to choose between more or less pelicans, I’d prefer more pelicans.
codethief 11 hours ago [-]
> Notably, Hy4 preview also contributed to its own development process, participating for the first time in the automated optimization of training methods, data strategies, evaluation frameworks, and low-level operators. The model proposed approaches, ran experiments, and iterated based on the results, with the resulting code, logs, and feedback feeding into subsequent rounds of exploration. This established an early-stage recursive self-improvement loop.
This reminds me of one of the predictions from https://ai-2027.com/ . Only that there it's "OpenBrain" doing this, not the Chinese. And the authors of that paper were also slightly wrong about "Mid 2026: China Wakes Up": China woke up already a while ago. And:
> But China is falling behind on AI algorithms due to their weaker models. The Chinese intelligence agencies—among the best in the world—double down on their plans to steal OpenBrain’s weights.
No need to steal anything, they have already caught up.
And then there's this prediction for February 2027:
> Officials are most interested in its cyberwarfare capabilities: Agent-2 is “only” a little worse than the best human hackers
I think we're past that point now, too…
bredren 10 hours ago [-]
If the distillation "attacks" created useful inputs to open weight models, ai-2027 was directionally correct that the Chinese would find ways to extract IP from western firms. (Scaled account creation and grinding outputs etc is not a dramatic story element as spies, though!)
Whether the distillation has constituted "attacks" or has or will meet the bar of "stealing" IP is not super interesting to me, though.
vlyan 3 hours ago [-]
the chutzpah of calling it an `attack` or `stealing` is super interesting tho.
judge2020 8 hours ago [-]
Don’t need a “better” hacker if you have ten thousand AIs all trying literally every single possible thing to exploit a system with. The main issue is that this will eventually bring down the exploitation cost enough to target very minor targets who weren’t worth it before.
8 hours ago [-]
try-working 10 hours ago [-]
Just like how Windows 95 contributed to its own development process.
0xbadcafebee 10 hours ago [-]
The AI 2027 paper/website is exactly the same as random guesses from tech bros after a couple of beers telling you what they think the future will be. It has nothing to do with political theory, economic theory, game theory, or any other quasi-scientific or rigorous evaluation of real world events and predictable outcomes. It's just vibes. If they're wrong nobody will notice, if they're right people will call them geniuses.
minimaxir 13 hours ago [-]
Hy4 apparently has ludicrous traction on OpenRouter already (https://openrouter.ai/tencent/hy4-preview), with trillions of tokens processed in a couple days: more than GLM 5.3 in a week. That said, it's relatively cheap with a 5% cache cost when everyone is still doing 10%/20% cache costs, so Hy4 may be more compelling.
martinald 12 hours ago [-]
I wrote about this a couple of weeks ago. It's actually often the biggest cost and it tends to be hidden away on most platforms!
Btw I still haven't came across any decent model that is <$0.01/MTok cache costs apart from deepseek thru their official API (even with the price increases).
Seems like a bit of an opportunity for someone to take - drop cache read costs significantly.
sieve 5 hours ago [-]
For me, an average long session results in about 200-300M cached input, 4-800K input, 2-400K output. Mostly the lower bound. Output depends on how much the model thinks.
There are two problems here:
- cache hit pricing (both Muse Spark 1.2 Contributor and MiMo 2.5 are around the $0.002-3/M mark)
- cache persistence time
Muse Spark drops the cache in less than 5m. MiMo keeps it around for at least an hour based on my experience with whoever is serving it for OpenCode. This difference itself will inflate bills massively.
A 500K token input repeatedly read by MS 1.2 for full input price 12 times an hour = $0.60. You would be expecting $0.012. So a 50x difference. Same thing on MiMo 2.5 is $0.018 because of longer cache times.
sourcecodeplz 2 hours ago [-]
even with the 5m cache, Muse Spark Contribs is still best bang for your buck for the intelligence you get.
it is basically the old dsv4-flash prices, but even more smart.
dakolli 11 hours ago [-]
That's because Deepseek invented the paradigm of prompt caching, they are the SOTA when it comes these techniques. Despite them open sourcing all their research, nobody beats them.
edit: I do wish openrouter would let you sort providers by Cache Hit % and Cache cost. These are the only things that matter to me at this point when choosing a provider.
Bolwin 11 hours ago [-]
Cache hit % on openrouter is not a good metric, it's mainly driven by openrouter's own provider juggling than the providers themselves
andai 10 hours ago [-]
OpenRouter randomizes which provider gets your request by default right? I think you have to pass a specific provider in the request to prevent that. (Or set up a preset or something.)
This behavior makes it so you don't benefit much from the caching, unless you pin it to a single provider.
Implicated 9 hours ago [-]
> OpenRouter randomizes which provider gets your request by default right?
I'm not sure it's wholey accurate to say they "randomize" the provider, rather my assumption based on usage is that it's something like cheapest-ish/responded to the request within some reasonable-ish time/etc algorithm that chooses the provider on each request - which seems, remarkably questionable in terms of optimizing for user experience or hidden user costs.
> This behavior makes it so you don't benefit much from the caching, unless you pin it to a single provider.
I so very much recommend this approach. My avenues that automate llm calls to openrouter are setup to make api reqs to openrouter to determine best price/response/etc and then pin the request to that (and, preferably, a fallback if there's reasonable difference between #1 and #2) provider for that session. Otherwise you're going to have a bad time.
I'd imagine this could make things interesting in cases where one provider is offering different quants than the others and openrouter is just swapping you back and forth on a long agentic session.
fc417fc802 8 hours ago [-]
> This behavior makes it so you don't benefit much from the caching
I don't believe this is correct? AFAIK once it routes you to a provider for a given conversation that choice is sticky unless you hit technical difficulties. (It's more complicated than that, they recently added named routing strategies that you can append to the model name.)
IMO the relevant metric is cache TTL which isn't typically published AFAIK.
RussianCow 2 hours ago [-]
This is very much NOT my experience in practice, even though it's how I would expect it to work. OpenRouter will happily bounce you between several providers (none of which have downtime) even within the same session. Requesting specific providers is the only way I've been able to hit a cache rate above 90%.
dakolli 11 hours ago [-]
This is not true, there isn't even a way to see a cache hit % model for a specific model, that wouldn't make any sense. You are confusing what I'm saying with cache cost, that has nothing to do with effective cache hit %. I'm talking about when you click on a specific provider for a specific model, you can scroll down on the view and see their cache hit % for that model [0].
These cache Hit % are accurate, I've done a ton of testing of this myself. The cache hit % is one of the most important metrics as far as estimating cost. There are many providers with cheap cache reads, but have an effective cache hit % of 30%, making their cheaper cache pricing meaningless compared to another provider who charges more but has a 85% cache hit percentage.
scroll down on the provider/model card and you'll see a field called cache hit %, its different for every provider/model.
I don't use routing on openrouter, I strictly use models with a single provider and no fallback, at least for use with harnesses its pretty dumb to route requests to multiple providers you are busting your cache every other request and increasing costs by 20-50%.
Implicated 9 hours ago [-]
I think you're arguing the same general point that the person you're responding to is. But you're saying he's not understanding - he understands that they report a cache hit % but you can't look at that public metric with any level of accuracy _because_ most people aren't pinning their providers and they _are_ getting juggled around which is bringing that metric down. That's not to say that specific providers might have issues or worse cache implementations - but it stands that if openrouter is juggling the requests back and forth by default then _that alone_ is breaking caches on those requests in huge numbers.
dakolli 8 hours ago [-]
I know what they're saying. Why would openrouter calculate it thay way lol. They obviously dont. Think for a sec, they arent idiots.
RussianCow 2 hours ago [-]
How else would you expect them to calculate it?
ralusek 10 hours ago [-]
> Cache hit %
I thought you had to actively manage caches, do you not?
minimaxir 11 hours ago [-]
You can click the table headers to sort Ascending/Descending.
dakolli 11 hours ago [-]
You can sort by the cost cache cose, but you cannot by cache hit %. You have to click on the provider and see what their cache hit % is. A provider could have a super low cache cost, but a 50% cache hit percentage, making the cheap price of cache read's meaningless.
andai 10 hours ago [-]
Wait, what does that number mean? I thought it always uses the cache price when the prefix matches.
dakolli 9 hours ago [-]
When the prefix matches a request sent to the same Providor. The thing is the TTL is different for each provider, some cache for 5 minutes some cache for 1hr. Its ideal to only use one provider per agent session / and per model with the best cache hit % if you care about costs.
orbital-decay 10 hours ago [-]
>Deepseek invented the paradigm of prompt caching
Caching was always here, you don't need to do anything special to get it on a single user local backend running a base model or a chatbot in the first place. Among commercial providers, OpenAI adopted it in 4o first.
joegibbs 8 hours ago [-]
If you’re Tencent you can just plug it into some field somewhere that lots of people see right? Like how Meta could put their model on Instagram search
redox99 10 hours ago [-]
It's very likely tencent games those stats, buying their own tokens.
Dinux 12 hours ago [-]
Which explains why almost none of my request go though
npn 13 hours ago [-]
[flagged]
cyanydeez 13 hours ago [-]
i'd be curious if openrouter is just being gamed by these publishers by paying for the exposure.
wouldn't trust they dont do Capitalism like the rest of the AI field.
drob518 12 hours ago [-]
Of course they are. Of course they do. Nobody should be surprised by this.
tokai 12 hours ago [-]
>dont do Capitalism like the rest of the AI field
Like lobbying the US president to harm their competitors?
realo 12 hours ago [-]
I would suggest "lobbying" is not the correct word to describe all the corruption going on in the current USA administration cesspool.
noir_lord 11 hours ago [-]
lobbying/legalised bribery hard to say where one ends and another begins at times.
CamperBob2 10 hours ago [-]
Well, it sure as hell isn't capitalism.
blackqueeriroh 8 hours ago [-]
Lmao that’s exactly capitalism
CamperBob2 7 hours ago [-]
Where in the Wealth of Nations does a Trump appear?
andrekandre 7 hours ago [-]
i mean, theres capitalism as the ideal, and there is capitalism in practice, so maybe you are both right...
jamienk 8 hours ago [-]
Genuine Q about word optimization/token density:
If we create a stripped-down vocabulary with greater token density to use less resources and to resolve ambiguities earlier in the semantic process, aren't we creating NEWSPEAK and dragging along the worst aspects of it? The ambiguity and multi-valence of words is what creates more connections between words, increases the directionality of associations, and expands the potential subtlety and depth of meaning. By paring down (or requiring verifiability) we make it harder to say certain things, or at least make it harder to unintentionally say something that makes MORE or DEEPER sense than what we intended. If the token density becomes extreme, you're left with something like a calculator.
Maybe this is the ultimate path toward better coding? But the worse path toward better genuine thinking?
nbush 7 hours ago [-]
This is one of the dangers. AI boosters would say that humans already do this compression and it was accelerated by mass media and then the internet, and that model memory + context can be broad enough that compared to human capabilities the opportunities for depth and variability are even greater. But I think we know which way this optimization usually goes. Even the notion of a "fine-tune for subtlety" is a contradiction.
dnautics 7 hours ago [-]
I don't think so. It's pretty clear that LLMs use the higher level layers for reasoning, so a bit of logorrhea very possibly enriches the result quality.
vatsachak 6 hours ago [-]
Nah reducing token length means that we're just reducing English down towards a programming language like a nice demi-glace
fastball 12 hours ago [-]
I wish model providers would stop committing chart crimes in their releases.
- if you're gonna order the rest of the bar chart by rank, order your model accordingly.
- if you're gonna highlight a winner in a table of benchmarks, don't highlight your entire model row in the table.
Etc etc
jimbob45 1 hours ago [-]
I wonder if this is being reinforced via LLM because they see every other modeler doing the same thing.
3 hours ago [-]
mirekrusin 11 hours ago [-]
Read websites through llm.
jorl17 12 hours ago [-]
I experimented with Hy3 for a project and was surprised with how good it was. I don't know if it's good for coding, but as a general purpose agentic model, it was only beaten by deepseek4-flash in our tests. It was so close to deepseek behaviour I kept thinking it must have been forked from it.
alexfortin 7 hours ago [-]
For the last few days I've been experimenting with the _free_ version of Hy3 offered by Opencode Go and I was also surprised to see how (relatively) good it is on coding tasks too.
The free quota from Opencode Go is also surprisingly generous, I perhaps hit limits one or two times and I've been using it _a lot_ for implementation tasks (using e.g. GLM-5.3-flash for working on specs and planning next steps).
realty_geek 1 hours ago [-]
Has anyone tried CodeBuddy? Is it worth trying out?
xeonax 2 hours ago [-]
We humans should adopt grug, instead of claudish. It seems simple to understand. And has this melancholical feeling
Zigurd 12 hours ago [-]
Is anyone here working on a problem for which current generation LLMs are inadequate, but that could possibly be solved by the next release of a first tier LLM?
Or is it like bicycles? Unless your problem is named Tadej, you don't need a $13,000 bike.
comex 11 hours ago [-]
My experience is that even Opus 5 still tends to write buggy or low-quality code and makes serious mistakes when analyzing code. It's a lot better than before but still not something I trust. I've had less experience with Fable since I can't use it at work; I hear it's a step up but still has its limits.
For large tasks like a web browser or a compiler, even expensive swarms of frontier LLMs have not been shown capable of producing codebases that actually work. (Anthropic built a C compiler with Opus 4.6 but it lacked optimizations and apparently hit a complexity wall.)
I also want to use LLMs for reverse engineering, but apparently it's pretty hit-or-miss, especially if you're forced to use open-source models to avoid restrictions.
Zigurd 10 hours ago [-]
This reply is particularly interesting to me because most of my experience with actually using LLMs to get work done is with coding agents. But I only have a fairly narrow set of experiences: two pretty large solo Flutter projects. I am currently really pleased with Gemini as a coding agent. It could improve, but I think improvements are going to come from marginal gains in the harness and training material so it can catch things like misconfigured permissions in platform specific areas.
It's also interesting because, while coding agents are important and are a notable success, they are never going to be a multi trillion dollar business. And are there any other domains where LLMs have such a large impact?
TiredOfLife 55 minutes ago [-]
Opus 5 is weird. It scores high on benchmarks, but it seems that majority of those who try to use it day to day hate it
RGS1811 12 hours ago [-]
For me personally, the answer is no. Fable is adequate to do basically anything I want to do. My perspective, broadly speaking, is that we've saturated most of the benchmarks because we've largely saturated our capacity to verify models' work at scale. What's left is context-bound verification, i.e. the problem of ensuring that output matches intent and ambiguities in prompting were resolved correctly. Further advances in autonomy do not make that latter verification problem easier. If anything they make it harder as the output per task becomes more complex and therefore more taxing for a human to verify.
The solution to that (to my mind) would be not a better model but a basic shift in architecture beyond the current paradigm and into a setup where agents have durable, plastic memories and undergo contextual individuation over time. But at that point agents start to become quasi-persons and not tools.
lopatin 11 hours ago [-]
I asked a current generation LLM to make me $1k a week and it hasn't so far.
spacebanana7 11 hours ago [-]
I want to be able to generate my own Simlilirian movie by dumping the content of a book into an LLM.
Both animated and live action results would be acceptable.
Unfortunately most existing LLMs lack the capability to maintain context across tens of thousands of frames.
Demiurge 11 hours ago [-]
That sounds like an interesting challenge. Have you seriously considered solving it? Because in about 10 seconds I came up with a process that should work, provided enough compute power. Simply model the traditional film making process by starting with a script, character stories. Design your world, then design the storyboard, and all the scenes. Create a list of all the visual elements that need to be replicated between all the scenes. Then you have to built prompts and reference art of the objects, faces, people. Make sure to do multiple takes of each scene, and have the vLLM critique and analyze the performances and technicalities. Should work?
I think, also, like in the traditional film makers career, this process should be built iteratively, start with a fast food commercial, then do a music video, then you can probably do a short film. Continue to improve the process, and one day I’m sure the LLM film studio can make you any movie you want, provided you have enough tokens.
andybak 11 hours ago [-]
I'm getting a Poe's Law feeling. I'm genuinely unsure about whether this post is a stone cold parody or not. I think I need to turn off the internet and go to bed.
EDIT: Your username doesn't help, either.
8 hours ago [-]
spacebanana7 2 hours ago [-]
> Have you seriously considered solving it?
Sort of, but I want it to be relatively low on human effort. I feel burned by spending lots of time in 2023 learning image generation pipelines (using control net etc) only for that to be rendered trivial by the next generation of LLMs.
This movie would be only for personal consumption and I’m okay with waiting for model improvements.
CuriouslyC 2 hours ago [-]
Control nets are still useful in Krea and H3. For images, Krea can usually get close enough to a reference that it's not a big deal, but for H3 conditioning makes a big difference over prompting for complex actions.
bsenftner 10 hours ago [-]
Nobody wants to watch such films, they want to muck with the filmmaker, the generation apparatus. That's the product, if there is one here, and absolutely not the 3 hour epic that's spit out with 4 variations to choose between. That's work. We'll have other LLMs pointlessly tell us which should be watched, we'll view a summary, and vote the Oscar on that.
aabdi 3 hours ago [-]
the problem is you have to make the AI watch the whole thing to make sure it works.
I've done this sort of with comfyui/same agent factory stuff, but the verification loop only works for models like fable as planner/writer, with gemini as verifier for like a very short movie. Sub 3-5 mins. After that you burn through million tokens.
Can't go too low fidelity audio/video or it craps out. Too long video and it loses consistency. Look at only snippets, it lacks global consistency, etc.
aforwardslash 10 hours ago [-]
On that topic, check higgsfield cinema studio 4; they already provide amazing tech for the cinematic experience, somewhat similar to what you are describing.
bsenftner 10 hours ago [-]
The results are boring. Not because the content is boring, but because you can so easily remix the results. Human curation is what creates value with these, not dumping and consuming. A personal perspective of a human being ups the respect, where the exact same sentences generated by an LLM carry no such value.
spacebanana7 2 hours ago [-]
Tolkien did the human creative work - I just want a movie adaptation that’s as honest to the original text as possible. Think translating the text into video.
finebalance 2 hours ago [-]
They are different mediums. While the underlying story might be Tolkien's, every frame is an artistic choice and while LLMs can make a choice is many situations, they are unable to 1) keep it coherent b) make it meaningful because art, to me and most, in an outcome of human experiences and thought, which by definition an llm cannot do.
Zigurd 10 hours ago [-]
I think this is the best and most realistic reply so far: the ability to do this is close enough, and things like AI music are hints that there is a business model for this. Maybe I'm just jaded about CGI effects in movies currently, but I think the fact that people except that kind of thing as entertainment means you might get away with a fully AI movie that people will pay for.
There are two more points in favor of this kind of AI movie project: there's zero chance that anyone would greenlight a Hollywood budget for the Silmarillion, and it is beyond human capability to write that screenplay.
RobotCaleb 10 hours ago [-]
How would a computer generated video be live action?
clipsy 7 hours ago [-]
Since live action results are acceptable, this is already possible with current day LLMs. Just instruct one to hire a writer, director, cast, and crew to make the movie.
Plus, the token costs involved should be pretty low! (Other costs may not be.)
ezst 11 hours ago [-]
I saw a laptop earlier in the train that I asked ChatGPT, Claude and Gemini what it was, providing a brand, screen size and ports description. Gemini could never figure it out, Claude and ChatGPT eventually did, after multiple rounds of indirection, giving completely wrong answers (there was a perfect match for the problem statement, they all explored alternatives first).
LLMs are (probably) amazing at things I don't care about, and still suck at the mundane stuff you would have the marketing tell you they excel at.
er4hn 11 hours ago [-]
I was given a picture cube, which is like a Rubik's cube but every side is a unique picture. It came scrambled and I don't have an original reference image. I like to take videos of it and give it to llms to solve. I call it my agi test because it hasn't been solved yet
aetherspawn 3 hours ago [-]
Interesting problem! I don’t think I could solve it myself honestly.
arjie 5 hours ago [-]
3d modeling to an STL a part compatible to a visible cable raceway still fails even if I let Claude Fable use me as a robot that measures with calipers.
vessenes 10 hours ago [-]
Yes. Most of us are, still. The frontier is currently both at expanding ‘common sense’ / non-cheating outcomes for imprecisely specified software (that’s all software), and at expanding autonomy - ability to work longer unsupervised with success, oh, and also at expanding outside contextual reasoning about what’s being built, as in “hmm, that doesn’t look right or make sense, let me explore that.”
hgoel 10 hours ago [-]
Scientific physics simulations - even the frontier models just engage in rationalization of obviously unphysical results instead of understanding the system. They have the rote knowledge but fail to apply it unless their hand is held through the process.
nextaccountic 3 hours ago [-]
Today's models can just write code to run the simulation instead
_factor 12 hours ago [-]
Hardware debugging and firmware details lead to thinking/testing loops on all but the frontier here.
pianopatrick 8 hours ago [-]
I think the tech analogy for frontier models is going to be super computers.
Super computers keep getting better but most people don't need them for most things.
eunos 5 hours ago [-]
2000's supercomputer is today's (highest end) smartphone performance tho
tekacs 11 hours ago [-]
Yes, lots – I think that folks will hopefully discover more of these as they scale up their ambition, now that LLMs make a lot of previously difficult things far easier.
jiggawatts 10 hours ago [-]
This is the exact same type of comment I heard about computer hardware upgrades for three decades in a row.
“Very few people actually require a Pentium workstation, a 486 is perfectly adequate for the majority”
The logical fallacy is taking an extant distribution of “product capability” that is priced to fit what the market will bear and assuming the “next upgrade” simply tacks on a little bit more to the right hand rail of that curve.
No!
It shifts the entire curve!
Everything for everyone gets better and the top 1% of the most demanding users will continue to pay the same-ish premium.
“Nothing” will change.
Look at it this way: you can buy a $200 laptop for your kid or a $20,000 Mac with an M5 Ultra processor.
BOTH are vastly more powerful than either a $200 PC or a $20,000 “workstation” from 20+ years ago.
The “budget” 5.5 Instant model beats o1 and o3 which were “pro” models at the time of their release!
Zigurd 10 hours ago [-]
Intel didn't just surf some natural wave of demand for higher power personal computers. Intel found new needs for powerful PCs, especially in gaming, and they put a lot of marketing and industry relations dollars behind PC gaming.
In other words. PC users didn't figure out that they could buy super powerful PCs and play games on them, that was a carefully managed market transition.
What is going to do the same for LLMs?
jiggawatts 9 hours ago [-]
> Intel found new needs for powerful PCs
It wasn't "Intel" that found new uses for PCs, it was everybody who found new uses for them. Billions of people and millions of companies found uses for "more computer power".
It was only the journalists with limited imaginations (and no industry experience) who struggled to come up with potential uses.
> carefully managed market transition.
You make it sound like a conspiracy! It wasn't. It was simple capitalist competition. If Intel hadn't improved their products, their competitors would have left them behind.
That very nearly happened ten years ago because Intel become stuck on the 14nm process and their products stagnated while Apple, ARM, and AMD lapped them repeatedly.
> What is going to do the same for LLMs?
Everybody.
Are you saying that unless you're "carefully managed" by some third-party, you could not find any use for "unlimited intelligence on tap"?
jml78 10 hours ago [-]
Infra as code and devops shit. Fable is there in general because things it doesn’t know I can point at documentation and have it do a reasonable job. Opus 5 sucks. If I don’t have fable quota, I drop to Opus 4.8 and hold its hand.
tokai 12 hours ago [-]
A spanish rock solved that problem for free.
wiether 3 hours ago [-]
Reading OP's analogy I was like "even him don't need this bike now..."
I feel bad for him as a human, but as a cycling fan I'm glad that we'll have an interesting WC in Canada
I get buy with very cheap models and actually using my brain, you don't need these SOTA models. China will definitely win this AI 'war'
kennywinker 11 hours ago [-]
If the models stay open, it seems like everybody but anthropic/openai wins. i literally can’t see a downside. We can post-train the models to know about tienanmen square.
vatsachak 10 hours ago [-]
I'm liking where LLMs are headed:
They can do the difficult small level optimization, the boring but tedious code but cannot be tasteful.
That means I'm more valuable and more productive. Good stuff
Flere-Imsaho 2 hours ago [-]
This is basically the conclusion the creator (DHH) of Ruby on Rails has come to:
It's all going to be who has the best and most tasteful ideas. Interesting times indeed.
handfuloflight 7 hours ago [-]
This guy gets it.
joshheitzman 9 hours ago [-]
Maybe's its a problem with the hosting at novita.ai but I didn't got much useful out of this model as a coding agent.
coder543 9 hours ago [-]
Novita does not offer Hy4-preview on either OpenRouter or their own model list. Maybe you confused it with Hy3.
joshheitzman 6 hours ago [-]
You are correct.
bobby_coder_55 8 hours ago [-]
Unfortunately codebuddy login is not working for me in the United States of America
XCSme 11 hours ago [-]
I tried benchmarking it, but it keeps timing out/rate limiting, so the current provider(s) are unusable.
yipinwong 9 hours ago [-]
I am going to bring up graph issue for everyone of these announcements.
They all suck.
They shoulda put their stick where they belong, not at far left.
It just makes comparison to Deepseek 90% of them time as Hy4 has nothing to show off.
vcryan 13 hours ago [-]
I used Hy3 quite a bit for the type of tasks it was suited for. Excited about this. My one concern over Hy3 was speed. In theory, it could be served much faster as a smaller model but it was relatively slow everywhere I could get it (including from Tencent directly) but also several other inference providers.
Topfi 12 hours ago [-]
In my evals, I saw an unprecedented jump between preview and final release on Hy3, from unusable to competitive. Did you see similar in preview vs release version?
vcryan 12 hours ago [-]
Oh yes! I forgot about that. Yes, you can see this in benchmarks about hy3 preview and hy3 release still today because they measured them separately - it was significant.
ls612 9 hours ago [-]
Open Weights is where the action is at in the past couple months, I’d have to think the US frontier labs are getting nervous. Like Anthropic hasn’t released anything pushing the frontier since “the event” earlier this summer.
usernomdeguerre 12 hours ago [-]
is it just me or are the bar charts in the blog post strange? Higher numbers don't seem to correspond correctly to their actual height?
pixelesque 11 hours ago [-]
Looks okay to me.
The first column has both the Hy4 and Hy3 scores overlaid on one another (Hy4 is darker blue and the taller one), with both scores written below the top of the respective bar - maybe you're seeing that?
alanfranz 12 hours ago [-]
Probably AI generated.
But, what bars are clearly off? I couldn't spot any.
feynmanquest 12 hours ago [-]
Noticed that as well
throaway2525634 10 hours ago [-]
I, for one, welcome our new Chinese overlords.
Yash16 2 hours ago [-]
[dead]
pilotcat 9 hours ago [-]
[flagged]
onesandofgrain 12 hours ago [-]
[flagged]
yogthos 12 hours ago [-]
Seriously, without China we'd just have two parasitic companies hoarding this tech and deciding whom and how is allowed to use it.
Kuyawa 10 hours ago [-]
Claude and OpenAi are not allowed in Venezuela, so I thank China too and I swear to god I'll never use them and will be rooting for chinese models forever
aforwardslash 10 hours ago [-]
Im not particularly fan of the chinese, but no chinese model asked for my citizen card yet to complete a task. And apparently no chinese provider uses persona to manage this kyc information. OpenAI does, in EU space. Just saying.
unethical_ban 10 hours ago [-]
Vague and without substance. It easily passes as sarcasm, which means criticism but without any commentary, else it is sincere... but doesn't have any commentary.
087532379864 10 hours ago [-]
[flagged]
andsoitis 9 hours ago [-]
> open-sources
link to source code?
sezaidemirer 11 hours ago [-]
Congratulations, it turned out great!
zem 5 hours ago [-]
I was briefly impressed that https://hylang.org/ had released a 4.0 version!
snthpy 2 hours ago [-]
Me too!
petcat 12 hours ago [-]
> Tencent has released and open-sourced Tencent Hy4 preview, a next-generation large language model with 770B total parameters and 49B active parameters, and a context window exceeding 1M tokens.
There are no open source models, at least not useful ones (yet) [0]. Open weight is not the same as open source. The current "open weight" models are just opaque binary blobs you can run on your own computer instead of through a web API.
Imagine thinking that running a Photoshop binary on your own computer instead of through a SaaS web app means that it's "open source". Of course you think that's ridiculous.
mirekrusin 11 hours ago [-]
You can open source dataset without all the details how it was assembled.
Models are lossy compressed datasets you can pick up and amend (fine tune / continue training / alter) according to license they were released under.
Hy4 is released under OSI approved Apache License 2.0.
kennywinker 11 hours ago [-]
Parent poster is technically right - open “source” implies the source used to make something is open. The model source is training data and code, not just weights.
But the reality is, the weights are a useful artifact that you can use to create derivative works. So, dismissing it as a photoshop binary is as technically wrong as calling it open source.
Alpha3031 1 hours ago [-]
IIRC Nvidia claims to release enough data that it should be possible to fully reproduce Nemotron, so even if it's not as good as the current best models, GPT 5.1 or Opus 4.1.was still useful right? I guess it depends on what you wanted to do with them.
villish 10 hours ago [-]
Countries that aren’t competitive need access to training datasets so that they may train their own similarly capable models and be sure of the inputs. Governments cannot blindly trust open weight models from China and the US.
LtWorf 10 hours ago [-]
So windows is open source because the binaries are a lossy compression of the original source?
NitpickLawyer 4 hours ago [-]
Weights are not binary. A model is created at init time, with random values. After that, it is being modified using data. The key point is that the labs modify the models "as weights". That means that weights are the intended / preferred way of modifying a model. Which, coincidentally, matches the definition of source in Apache 2.0. There is no "higher level" place where editing takes place. It all happens in weight space. Through the license you get the same rights as the lab that created it: view, inspect, run, modify, re-release. That's it. That's the only thing a license can grant you.
The rest is semantics, misunderstandings, and FUD. A model released under an open source license is open source. Training data is lab knowhow / IP. Which, historically, has never been required for any open source release.
frabcus 2 hours ago [-]
Well, you can't add or alter data in pre-training from just the weights. Which, as I understand it, means you can't fundamentally increase core knowledge or cognitive ability, only what the model likes to do with those. You can only post-train, and you're subject as a result to catastrophic forgetting.
To explain simply as far as I can tell (would love to be corrected) the large number of pre-training tokens only works because the documents are randomly ordered.
So if you e.g. took a foundation model with open weights, then tried post-training it all the new data since its cut-off period, it would then end up over-trained on that new data, and forget older things.
mirekrusin 2 hours ago [-]
As I live next to EPFL, I'll give you example from them: their Meditron-70B model is adapted to the medical domain from Llama-2-70B through continued pretraining. They took weights of Llama-2-70B and continued training on PubMed, medical guidelines and general data.
Weights aren't just executable artifact that's consumed by users. Third parties actually use released parameter state as the editable starting point for further training and produce new foundation models from it.
> Maybe add sunglasses? no.
> Maybe add water? no.
https://tools.simonwillison.net/markdown-svg-renderer#url=ht...
Edit: someone else commented that as I was typing this, lol.
Maybe you should start also comparing reasoning traces when you do your pelican benchmark.
https://huggingface.co/ProCreations/grug-27b
https://github.com/p-e-w/heretic
Fine-tuning is great for really small models on specific applications, but it's not something that can essentially improve a more generic model.
That said, there seems to be a fine line in quantization+finetuning that could recover performance. It's just hard to get a hold of it (I feel it in some models, but it's hard to say yet; lots of small labs working on this RN).
This is also likely to stop working as censoring moves to the training data source.
Be concise.
Brief is best. Eschew verbosity-- William Strunk Jr. and E.B. White., The Elements of Style
Real humans get non-primary information from word variation. It's reasonable to hypothesize that it has a role in thinking things, because it endures. Our languages need to breathe over time, and flourishing might be one of the aspects that allows that breathing space.
Maybe add a small cycling cap or helmet if it doesn’t obscure the head.
Chinese can be extremely information-dense in token terms, though it depends on the tokenizer. Roughly speaking, you can pack more "meaning" into a short sequence than English often allows for. That's why "caveman" reasoning is a pretty good fit.
There's a difference between bolting caveman speak onto an existing model and training a model to reason that way, though. If you just force an existing model to be concise in outputs, you're artificially reducing its available reasoning steps and can possibly prevent useful exploration or verification. If it's trained specifically to use compressed reasoning, it can learn to represent the same intermediate ideas in fewer generated tokens, cutting the number of sequential inference steps without necessarily sacrificing the useful reasoning itself.
It's not so much inherently a Chinese-model trait, but Chinese models could definitely have helped demonstrate how effective very compressed reasoning traces can be.
There are few tests of this, but one example I thought was interesting was here: https://github.com/PastaPastaPasta/llm-chinese-english
I wouldn't say it was Chinese specifically that was emulated, but it got people thinking about tokenizers and representation efficiency, and how natural English is rather inefficient.
A "train of thought" can be seen as a trace of a depth first search where the preceding trace is used to guide termination and next expansion decisions. A similar concept, "taboo search", exists in classical constraint optimization where previous solutions are fit to a model that guides future expansion (but as the name "taboo" implies, away from uninteresting solutions).
We also have harnesses that perform breath first search.
If I tried to describe what it means to "think deeply", I would probably say a combination of both.
Ultimately I believe that we will surpass human capabilities but fail with alignment. Handing the world's resources over to stochastic systems that can evolve faster than we can reason about them simply leaves too many "interesting" outcomes that do not end well. I also expect the failure modes will be totally non-obvious.
I think this statement is continuation of the old fallacy - every generation thinks of brain in terms of what is the current technology zaitgeist is - was it 19th century when they thought brain is a network of pneumatic pipes?
That’s the reason we are “doing fine”. Once they stop needing you..
Also, both our comments brush over the generational struggles for fairness over the centuries. We have fought to be “fine”, it did not just happen. Without fairness being introduced by force you and I would be slaving away in some sweatshop getting paid nickels as was the norm not so long ago.
Edit: That’s also assuming you are Caucasian. If you are of a different ethnicity.. well, historically, all bets are off. You could also be the literal possession of some of these “worst people” with not even your own children considered yours.
Hell no. In very narrow tasks - yes, in vast majority, esp. involving state tracking (board games) and spatial reasoning - they are awful.
The other option is that you do understand those words the same way, and the people making these (now nonsensical) anti-AI claims simply aren’t talking about the same programs/models we are. Their idea of SOTA is when chatgpt.com launched.
If you took a point sample pre-Opus, and didn’t write a good prompt, of course you would think all AI programming was worthless slop.
Just tap on the [-], and upvote what you find more interesting :)
This reminds me of one of the predictions from https://ai-2027.com/ . Only that there it's "OpenBrain" doing this, not the Chinese. And the authors of that paper were also slightly wrong about "Mid 2026: China Wakes Up": China woke up already a while ago. And:
> But China is falling behind on AI algorithms due to their weaker models. The Chinese intelligence agencies—among the best in the world—double down on their plans to steal OpenBrain’s weights.
No need to steal anything, they have already caught up.
And then there's this prediction for February 2027:
> Officials are most interested in its cyberwarfare capabilities: Agent-2 is “only” a little worse than the best human hackers
I think we're past that point now, too…
Whether the distillation has constituted "attacks" or has or will meet the bar of "stealing" IP is not super interesting to me, though.
https://martinalderson.com/posts/watch-out-for-cache-read-co...
Btw I still haven't came across any decent model that is <$0.01/MTok cache costs apart from deepseek thru their official API (even with the price increases).
Seems like a bit of an opportunity for someone to take - drop cache read costs significantly.
There are two problems here:
- cache hit pricing (both Muse Spark 1.2 Contributor and MiMo 2.5 are around the $0.002-3/M mark)
- cache persistence time
Muse Spark drops the cache in less than 5m. MiMo keeps it around for at least an hour based on my experience with whoever is serving it for OpenCode. This difference itself will inflate bills massively.
A 500K token input repeatedly read by MS 1.2 for full input price 12 times an hour = $0.60. You would be expecting $0.012. So a 50x difference. Same thing on MiMo 2.5 is $0.018 because of longer cache times.
it is basically the old dsv4-flash prices, but even more smart.
edit: I do wish openrouter would let you sort providers by Cache Hit % and Cache cost. These are the only things that matter to me at this point when choosing a provider.
This behavior makes it so you don't benefit much from the caching, unless you pin it to a single provider.
I'm not sure it's wholey accurate to say they "randomize" the provider, rather my assumption based on usage is that it's something like cheapest-ish/responded to the request within some reasonable-ish time/etc algorithm that chooses the provider on each request - which seems, remarkably questionable in terms of optimizing for user experience or hidden user costs.
> This behavior makes it so you don't benefit much from the caching, unless you pin it to a single provider.
I so very much recommend this approach. My avenues that automate llm calls to openrouter are setup to make api reqs to openrouter to determine best price/response/etc and then pin the request to that (and, preferably, a fallback if there's reasonable difference between #1 and #2) provider for that session. Otherwise you're going to have a bad time.
I'd imagine this could make things interesting in cases where one provider is offering different quants than the others and openrouter is just swapping you back and forth on a long agentic session.
I don't believe this is correct? AFAIK once it routes you to a provider for a given conversation that choice is sticky unless you hit technical difficulties. (It's more complicated than that, they recently added named routing strategies that you can append to the model name.)
IMO the relevant metric is cache TTL which isn't typically published AFAIK.
These cache Hit % are accurate, I've done a ton of testing of this myself. The cache hit % is one of the most important metrics as far as estimating cost. There are many providers with cheap cache reads, but have an effective cache hit % of 30%, making their cheaper cache pricing meaningless compared to another provider who charges more but has a 85% cache hit percentage.
[0]: https://openrouter.ai/deepseek/deepseek-v4-flash-0731?endpoi...
scroll down on the provider/model card and you'll see a field called cache hit %, its different for every provider/model.
I don't use routing on openrouter, I strictly use models with a single provider and no fallback, at least for use with harnesses its pretty dumb to route requests to multiple providers you are busting your cache every other request and increasing costs by 20-50%.
I thought you had to actively manage caches, do you not?
Caching was always here, you don't need to do anything special to get it on a single user local backend running a base model or a chatbot in the first place. Among commercial providers, OpenAI adopted it in 4o first.
wouldn't trust they dont do Capitalism like the rest of the AI field.
Like lobbying the US president to harm their competitors?
If we create a stripped-down vocabulary with greater token density to use less resources and to resolve ambiguities earlier in the semantic process, aren't we creating NEWSPEAK and dragging along the worst aspects of it? The ambiguity and multi-valence of words is what creates more connections between words, increases the directionality of associations, and expands the potential subtlety and depth of meaning. By paring down (or requiring verifiability) we make it harder to say certain things, or at least make it harder to unintentionally say something that makes MORE or DEEPER sense than what we intended. If the token density becomes extreme, you're left with something like a calculator.
Maybe this is the ultimate path toward better coding? But the worse path toward better genuine thinking?
- if you're gonna order the rest of the bar chart by rank, order your model accordingly.
- if you're gonna highlight a winner in a table of benchmarks, don't highlight your entire model row in the table.
Etc etc
The free quota from Opencode Go is also surprisingly generous, I perhaps hit limits one or two times and I've been using it _a lot_ for implementation tasks (using e.g. GLM-5.3-flash for working on specs and planning next steps).
Or is it like bicycles? Unless your problem is named Tadej, you don't need a $13,000 bike.
For large tasks like a web browser or a compiler, even expensive swarms of frontier LLMs have not been shown capable of producing codebases that actually work. (Anthropic built a C compiler with Opus 4.6 but it lacked optimizations and apparently hit a complexity wall.)
I also want to use LLMs for reverse engineering, but apparently it's pretty hit-or-miss, especially if you're forced to use open-source models to avoid restrictions.
It's also interesting because, while coding agents are important and are a notable success, they are never going to be a multi trillion dollar business. And are there any other domains where LLMs have such a large impact?
The solution to that (to my mind) would be not a better model but a basic shift in architecture beyond the current paradigm and into a setup where agents have durable, plastic memories and undergo contextual individuation over time. But at that point agents start to become quasi-persons and not tools.
Both animated and live action results would be acceptable.
Unfortunately most existing LLMs lack the capability to maintain context across tens of thousands of frames.
I think, also, like in the traditional film makers career, this process should be built iteratively, start with a fast food commercial, then do a music video, then you can probably do a short film. Continue to improve the process, and one day I’m sure the LLM film studio can make you any movie you want, provided you have enough tokens.
EDIT: Your username doesn't help, either.
Sort of, but I want it to be relatively low on human effort. I feel burned by spending lots of time in 2023 learning image generation pipelines (using control net etc) only for that to be rendered trivial by the next generation of LLMs.
This movie would be only for personal consumption and I’m okay with waiting for model improvements.
I've done this sort of with comfyui/same agent factory stuff, but the verification loop only works for models like fable as planner/writer, with gemini as verifier for like a very short movie. Sub 3-5 mins. After that you burn through million tokens.
Can't go too low fidelity audio/video or it craps out. Too long video and it loses consistency. Look at only snippets, it lacks global consistency, etc.
There are two more points in favor of this kind of AI movie project: there's zero chance that anyone would greenlight a Hollywood budget for the Silmarillion, and it is beyond human capability to write that screenplay.
Plus, the token costs involved should be pretty low! (Other costs may not be.)
Super computers keep getting better but most people don't need them for most things.
“Very few people actually require a Pentium workstation, a 486 is perfectly adequate for the majority”
The logical fallacy is taking an extant distribution of “product capability” that is priced to fit what the market will bear and assuming the “next upgrade” simply tacks on a little bit more to the right hand rail of that curve.
No!
It shifts the entire curve!
Everything for everyone gets better and the top 1% of the most demanding users will continue to pay the same-ish premium.
“Nothing” will change.
Look at it this way: you can buy a $200 laptop for your kid or a $20,000 Mac with an M5 Ultra processor.
BOTH are vastly more powerful than either a $200 PC or a $20,000 “workstation” from 20+ years ago.
Look at: https://arena.ai/leaderboard/text?q=openai&utm_source=chatgp...
The “budget” 5.5 Instant model beats o1 and o3 which were “pro” models at the time of their release!
In other words. PC users didn't figure out that they could buy super powerful PCs and play games on them, that was a carefully managed market transition.
What is going to do the same for LLMs?
It wasn't "Intel" that found new uses for PCs, it was everybody who found new uses for them. Billions of people and millions of companies found uses for "more computer power".
It was only the journalists with limited imaginations (and no industry experience) who struggled to come up with potential uses.
> carefully managed market transition.
You make it sound like a conspiracy! It wasn't. It was simple capitalist competition. If Intel hadn't improved their products, their competitors would have left them behind.
That very nearly happened ten years ago because Intel become stuck on the 14nm process and their products stagnated while Apple, ARM, and AMD lapped them repeatedly.
> What is going to do the same for LLMs?
Everybody.
Are you saying that unless you're "carefully managed" by some third-party, you could not find any use for "unlimited intelligence on tap"?
I feel bad for him as a human, but as a cycling fan I'm glad that we'll have an interesting WC in Canada
They can do the difficult small level optimization, the boring but tedious code but cannot be tasteful.
That means I'm more valuable and more productive. Good stuff
https://lexfridman.com/dhh-david-heinemeier-hansson-transcri...
It's all going to be who has the best and most tasteful ideas. Interesting times indeed.
They all suck.
They shoulda put their stick where they belong, not at far left.
It just makes comparison to Deepseek 90% of them time as Hy4 has nothing to show off.
The first column has both the Hy4 and Hy3 scores overlaid on one another (Hy4 is darker blue and the taller one), with both scores written below the top of the respective bar - maybe you're seeing that?
But, what bars are clearly off? I couldn't spot any.
link to source code?
There are no open source models, at least not useful ones (yet) [0]. Open weight is not the same as open source. The current "open weight" models are just opaque binary blobs you can run on your own computer instead of through a web API.
[0] https://allenai.org/
Imagine thinking that running a Photoshop binary on your own computer instead of through a SaaS web app means that it's "open source". Of course you think that's ridiculous.
Models are lossy compressed datasets you can pick up and amend (fine tune / continue training / alter) according to license they were released under.
Hy4 is released under OSI approved Apache License 2.0.
But the reality is, the weights are a useful artifact that you can use to create derivative works. So, dismissing it as a photoshop binary is as technically wrong as calling it open source.
The rest is semantics, misunderstandings, and FUD. A model released under an open source license is open source. Training data is lab knowhow / IP. Which, historically, has never been required for any open source release.
To explain simply as far as I can tell (would love to be corrected) the large number of pre-training tokens only works because the documents are randomly ordered.
So if you e.g. took a foundation model with open weights, then tried post-training it all the new data since its cut-off period, it would then end up over-trained on that new data, and forget older things.
Weights aren't just executable artifact that's consumed by users. Third parties actually use released parameter state as the editable starting point for further training and produce new foundation models from it.