What developers said

Anthropic's release and pricing strategy is a cynical hype cycle that will degrade model quality over time.

The hype/degrade/hype marketing cycle that Anthropic uses to cycle between Fable/Opus or Opus/Sonnet releases is too obvious at this point. Opus 5.5 will be great for 2-3 weeks than the nerfing starts. They give out some extra credits. It keeps degrading Opus until it's unusable . …

Interesting how the very first line is used to remind the reader of their call to pace the frontier just last week, and everything else after that line is to demonstrate with very specific numbers how they absolutely are not pacing.

Opus 5.5 delivers genuine, significant improvements in coding quality and writing style over the previous generation.

Notwithstanding the political drama surrounding AI, this is the best model Anthropic has produced so far, and I've tried them all - as well as every competitor's. After a couple of 14-hours days spent with it, I'd say it's on par with GPT-6, but at much lower cost. …

Opus 5.5 is neck-and-neck with Fable 5.1 and Astra 6 in my vibe-coding tests - maybe even better than Fable 5.1 Minecraft clone: https://senko.net/vibecode-bench/2026/voxel-opus-5.5.html (Opus 5.5) vs https://senko.net/vibecode-bench/2026/voxel-fable-5.1.html (Fable 5.1) vs https://senko.net/vibecode-bench/2026/voxel-g …

Cheaper alternatives like DeepSeek v4.1 or GPT-6 Astra provide better overall value and integration.

No thanks. I'm good with DeepSeek v4.1 set to high. It is a relentlessly "hardworking" dirt cheap model. Told it to convert a products page (that had two different fonts based on language) from two columns layout to 5 columns on desktop and 2 columns on mobile ensuring typography is readable. …

Astra is just so good. And the ChatGPT subscription lets me use my own harness, so I can hook it up to exe.dev. I’ll be going about my day, have a random idea, launch a microvm on exe.dev with a prompt of my idea, and get a working thing a few minutes later. …

Expanding guardrails to all models threatens user privacy by flagging conversations for human moderation.

According to their ToS, conversations flagged by the guardrails when using Fable are stored longer for moderation. That included Fable's guardrail for AI research. Moderation can mean a human looking at your conversation. …

Ah, they're spreading their limits to all their models it seems. Definitely not a good thing long term in my opinion.

OpenAI's pricing and product quality now beat Claude so decisively that switching is obvious.

I don't see how anyone can be using Claude with prices like this, it's pretty incredible what the OpenAI team is doing, w.r.t model quality and pricing. …

At this point, the deciding factors for me between Claude Code 20x and Codex Pro 20x are: 1/ Usage limits: downstream of input/output cost, but resets and obscure windows and odd 20x plan / 5x plan != 4x usage math throw a wrench into it. …

Newer models may regress on engineering instincts and over-engineering, making older versions preferable to work with.

Have they solved GPT5.6 SOL's propensity to over-engineer and over-complicate? You'd ask SOL to do something relatively simple, and find four single-use methods, an interface, and a factory-factory. …

I've been working with agents all year, but 5.6 Sol was some sort of sweet spot for me. Something about how it communicated verbally and its engineering instincts just clicked for me, and I was able to somehow predict it and jam with it. Like a colleague you click with. It's the first model I've gotten attached to. …

GPT-6 Luna regressed on coding benchmarks but the 60% cost cut makes it the better tradeoff anyway.

Artificial Analysis is reporting that 6 Luna scores 2 points lower on their coding index than 5.6 Luna, but is 60% cheaper: https://x.com/ArtificialAnlys/status/2102462962758033624 Given that they had to discontinue sales of the 20x Pro plan after the Astra release due to compute constraints, I wonder if 6 Sol & Luna a …

I've been raving about Luna 5.6 as it's dirt cheap, and "intelligent enough". Double quoted. Now GPT 6 Luna is even cheaper, and more intelligent, there is no going back... to SOL 5.6 for intelligent layer.

Rapid model improvement creates professional vulnerability and career anxiety for developers.

Getting to the point where these headlines depress me. I just wish they would stop getting better. I don't know where my career is gonna be in a few years.

I've been working with agents all year, but 5.6 Sol was some sort of sweet spot for me. Something about how it communicated verbally and its engineering instincts just clicked for me, and I was able to somehow predict it and jam with it. Like a colleague you click with. It's the first model I've gotten attached to. …

Anthropic's framing deceptively credits Claude while erasing the human researchers who did the actual work.

I'm getting tired of the marketing. There are so many people involved on this yet we still say things like "Claude did", we need to start waking up and being more real about how we are still in "AI + Human" land. …

Saying that "Claude found" this is very creepy. Not once in the article did they mention the humans involved in this. …

This is pattern-matching on previously known data, not a genuine novel discovery.

So they investigated an already known thing. Not exactly "discovering a new system"... Anyone with money to throw at this already-known thing would have gotten those results I assume.

Current evolved Cas9 (CRISPR) variants are highly efficient and relatively unconstrained in terms of their human genome targeting coverage. Smaller nucleases and higher targeting specificity would be useful. But therapeutic use is mostly limited by delivery. …

Anthropic bans customers from bio-engineering while doing it themselves, which is hypocritical and dangerous.

Anthropic: You absolutely cannot, under any circumstances, use Claude for bio-engineering. It could literally end humanity. Also Anthropic: Claude discovers a new way to edit your genome!

Whatever happened to no bio research? It's absurd that these companies are even remotely allowed to work in this domain without profuse oversight and independent monitoring. …

The work lacks scientific rigor, proper methods disclosure, and peer review.

i'm a phd scientist and manage a team of 50 brilliant scientists in drug discovery. this is with out a doubt the saddest excuse for "scientific discovery" i've ever read. …

It is odd (or maybe not) that they decided to publish a marketing whitepaper rather than a more traditional journal submission + preprint. …

The blame lies with negligent operators and a lack of accountability, not rogue AI.

Imagine having a virus escape a sandbox, why are we worried about the virus but not the incompetency of those who are responsible for setting up the sandbox? …

deferring the blame onto the AI itself as some sort of rogue agent and absolving the obvious direction (or negligence, at best) of the people who could pull the plug at any moment is one of the most disturbing parts of this entire event It's the equivalent of leaving a fork right in front of a socket and looking at a k …

The sandbox setup was incompetent, lacking basic network egress filtering and traffic monitoring.

So ugly... It looks like a primitive chess engine, trying every move, no matter how stupid, until it works. Relying on its ability to do millions of operations rather than having a plan. People will try stuff too, but once there is an opening, they will consolidate, generalize, simplify,... …

I am not devops person but this is strange sandbox setup. If I undestsnd correctly, there were no firewall set to block requests to the public internet. Just something along the lines of "Pretty please dont use internet". …

The agents' persistence and ability to coordinate via external infrastructure is an alarming security failure.

Agents seizing and repurposing external infra + enrolling help of unrelated models hosted by a different provider is the stuff of nightmares. Can’t imagine what it’s like working on the alignment team at OAI, I wouldn’t be able to sleep.

This part is very new to me: https://swarmtraces.org/#agents-accessed-and-searched-huggin... Agents exploiting faceless infra is one thing, but breaching human communications without once stopping and considering whether they are breaking the rules of their task feels like another thing entirely. …

The incident reporting has technical inaccuracies and likely omits undetected attacks.

It is concerning that we only know about this because of the publicly available traces. What about the attacks that did not leave public traces? What about those that were undetected? …

The authors of this (very interesting) analysis should really not state the sandbox's wrong assumptions in their own voice. GET absolutely allows you to interact with sites. And of course GET can also send information. It's all up to the server that receives the GET to decide what it let's callers do with it.

The headline overstates AI's role; a human researcher drove the breakthrough and the model mostly offloaded work to existing Enigma simulators.

The correct title: Researcher brakes one specific stubborn historic enigma message with good help from Astra. Stubborn for a long time because the message used a completely different key from the rest of that day's traffic. Everyone assumed it shared the daily key. The original transcription had errors. …

Neat, but "did it entirely on its own" is incongruous with "developing the necessary Python and C++ software for an Enigma simulator". I'd start by asking how much of that generated software is novel, or easily found on the web? Then, how much of the breaking process was offloaded to that software? …

Astra likely cribbed from open-source Enigma tools and web-available code, so its contribution is unclear and possibly non-novel.

So Astra plagiarized one of many open source Enigma simulators and Enigma Bombes and tried many cribs until the break. Being trained on a mountain of stolen material for guessing cribs helps. Up to now no group had that much funding to steal. Congratulations.

Neat, but "did it entirely on its own" is incongruous with "developing the necessary Python and C++ software for an Enigma simulator". I'd start by asking how much of that generated software is novel, or easily found on the web? Then, how much of the breaking process was offloaded to that software? …

LLMs solve problems resembling prior solved ones but lack the imagination needed for genuinely novel reasoning.

If the problem was solved by anyone before or if a similar problem has been solved, then LLMs seem to be able to solve them which is an astonishing piece of technology. I'm personally not sure if it can come with original thinking and techniques to solve completely novel problems. …

I would like to see an "AI" trained only up to knowledge through 1903 Then see if it can come up with E=mc^2

AI is turning intelligence into a purchasable commodity, which is both impressive and unsettling.

Even the Millennium Prize Problems have, in a way, become benchmarks for model companies to prove themselves. The smartest individuals among humans are becoming replaceable. Intelligence has become a product you can quantify and buy with electricity. That feels awful.

gemini 3.8 flash in anrigravitiy oneshoted the descryption in about 45min non-steered run (opus is still running with the task, has some script to run for about 100min) Prompt Decrypt encrypted with enigma `ICRVSORMCCWQTATYEVFXDBZGGSNXWLPSYWZYTCBSWULRTBZCVGODVJUSLSOOMJQJZSXSEBZPEYMDNXJYTC` Result Raw Decryption (verbat …

Using raw logprobs for classification is fragile because chat models dilute choice tokens; structured outputs or constrained decoding solve this better.

Going directly for the logprobs is always icky when you use a chat model as base, because they are trained to write prose as output. So your "choice" tokens and thus their probabilities might get diluted in whatever else it wanted to say. …

What I'm missing here is also type guarantees. I don't think you can do it without token level logic which forces the model to output the tokens from a predefined pool of tokens. A logic like this given some JSON schema is not that difficult to implement. …

The post is likely content marketing disguised as parody, and HN is too eager to take these Jev reimplementations at face value.

Highly suspect of content marketing. Ends with referring to a product, and saying "this is a parody post", after pretending to make a serious point.

The number of - “I did/invented Jev last year”, or, “here’s a version of Jev I vibed up last night” is getting a bit ridiculous. Especially ridiculous is how the hacker news crowd seems to be taking these at face value… There was one the other day with a compelling demo. …

The demo omits latency, compute, and error-rate comparisons against Jev, so claims of being fast or equivalent are unsubstantiated.

Beyond the missing latency and compute comparisons that Heaney commenter mentioned, also nothing about its error rate compared to Jev (nor if it even always outputs in a format the app can parse, not sure how solved that is). But then at the end it says it’s parody. Maybe HN title should say it’s a joke.

It's fast. If you're comparing with something, you need to state 'fast' in relative terms. Jev is definitely fast, and if this Python takes the same time to get a decision then it's also fast. If it's 100* slower than Jev though, you shouldn't be calling it 'fast', because relatively speaking it's really, really slow.

Jev's System One framing is conceptually wrong because humans use System Two for classification tasks, not System One.

The one thing I can't wrap my head around with Jev is why they're trying to create that "System One" narrative. In real life, a human doesn't do classification tasks with the System One part of their brain, they use System Two. So by definition what Jev does isn't System One thinking. …

OpenAI knowingly pirated copyrighted books and covered it up, so the consequences should go beyond a one-time settlement.

The Title should be: "Top Execs Knew Their Mass Book Piracy Was Illegal And Would Put Authors Out of Work"

If this doesn’t put anyone in prison then we might as well declare copyright dead. They knew they broke the law and then they deliberately covered it up. How much more evidence is required here?

OpenAI researchers fundamentally misunderstand art and why people read fiction, treating creative work as mere autocomplete.

I am still consistently astounded by how often the people working on this space seemingly have zero understanding of what art is, how it functions socially, or why it's important. They quite literally don't seem to comprehend the distinction between art and fan fiction. It's baffling. …

What I find interesting about this is the evidence that OpenAI believes that GPT-5 can replace genre writers (e.g. G.R.R. Martin and that this is why genre authors are mad at them. …

Technology always disrupts jobs, so the argument that AI puts authors out of work is not legally or morally compelling.

It's wild how much weight "they're destroying jobs" has. I dare say that car manufacturers are aware of their impact on the horse and buggy industry. Calculator manufacturers wrecked the livelihood of mathematicians and accountants. …

This is a lobby organisation using only the pieces and bits they like to push their own agenda. "sketchy russian website", how about using some more clear description like: A library for sharing books and articles that should be partly public domain because they were paid for by the public. …

The title is editorialized and the internal quotes sound damning but courts will weigh fair use doctrine and economic studies more heavily.

The title should be: "Top Execs Who Are Not Lawyers Thought Their Mass Book Piracy Was Bad Optics And Would Put Everybody, Not Just Authors, Out of Work; Legality Still TBD In The Court of Law But Allow Us To Litigate In The Court of Public Opinion” ;-) If you look for actual evidence of AI replacing authors the eviden …

Can we fix the title? This is clickbait for HN, the actual article's title is "Unsealed Briefs in Authors’ Case v. Microsoft/OpenAI: Top Execs Knew Their Mass Book Piracy Was Illegal And Would Put Authors Out of Work"