HN user
Couldn’t this effect be classic cause vs correlation?
Perhaps someone who has a consistent schedule is hypothetically more likely to make healthier choices on average?
Anthropic released their PPTX skill[0] a while ago, the relies a bit on python and pptxgenjs, depending on the usecase:
Read/analyze content....python -m markitdown presentation.pptx Edit or create from template....Read editing.md Create from scratch...Read pptxgenjs.md
[0] https://github.com/anthropics/skills/blob/main/skills/pptx/S...
Agreed. There’s also something about “time under consideration.” The relatively slow pace of reading the book opens the door to personal insight that a summary hides. The words themselves might be important, but I’ve found the additional minutes of a book is critical for me to fully apply the concepts to my life.
what's the desired outcome here?
I know agencies get paid to manage other business's instagram accounts, and want ways to post/comment/engage/etc with multiple business's accounts at once. You'd still have a human driving/scheduling/approving the activity, but you wouldn't have to jump between so many hoops to do so.
Can someone help me understand how its "confidential" if they blog about it? Perhaps they simply mean the details of the S-1 are confidential for now?
Agreed! OpenAI even bought TBPN [1], who many have equated to ESPN for business. I think that even if Karpathy didn't add any new ideas to Anthropic (unlikely), adding him to the team is an interesting message to give to the market
Honestly at $500 I still want one. I’d love to see the design open sourced!
Why? What do you mean by “bunk?” Do you have any examples you can point to?
Looks like the top downloaded book yesterday[0] was Concrete Construction: Methods and Costs by Gillette and Hill.[1] Beat out Moby Dick, Count of Monte Cristo, Frankenstien, Romeo and Juliet, and others.
23644 downloads in the last 30 days.
I wonder if this is bot behavior? 23k downloads feels like a lot?
[0] https://www.gutenberg.org/browse/scores/top [1] https://www.gutenberg.org/ebooks/24855
Yeah I totally agree with this list. In contrast, Coros had a pretty nonchalant response to their security issues last year. Attackers could:
Hijacking the vicitim’s COROS account and accessing all data
– Eavesdropping sensitive data, e.g. notifications
– Manipulating the device configuration
– Factory resetting the device
– Crashing the device
– Interrupting a running activity and forcing the recorded data to be lost [0]
The security firm disclosed the vulnerability to Coros in Mar 2025. They planned to fix it by the end of 2025, and didn’t address it until the security firm publicly released the finding.
[0] https://www.dcrainmaker.com/2025/06/coros-confirms-substanti...
Hey OP, would love to know more about your thoughts on Garmin you reference at the bottom? Why would they be any better/worse than Coros?
*= I do not like Garmin, I think they're a fraudulent company systematically breaching consumer rights and I'm looking for alternatives. Already converted multiple people to Coros.
I was using the term as you used in your comment. I believe the official term is "Activation" however:
Activation: When a task matches a skill’s description, the agent reads the full SKILL.md instructions into context.[1]
Full instructions load only when a task calls for them, so agents can keep many skills on hand with only a small context footprint.
The metadata is loaded by the harness, but the LLM still needs to choose to load the rest of the skill, no?
The example model in the documentation is 4o-mini, you might want to update that to a more recent model.
As an aside, 4o-mini came out months before agent skills were released… I’m curious how it performs with choosing to load skills in the first place?
I recently tried to sign up for a new domain with a .ai namespace. I tried around 50 names. All of them were not only taken, but seemingly have a landing page with varying degrees of functionality described.
Try it - type a word into your browser with .ai and you’ll see
For those unfamiliar with WebRTC, the Pion FAQ page has a good description:
WebRTC is a standardized protocol for P2P communication. It allows two peers to exchange media and data. It is encrypted by default, and handles connectivity establishment in many different network conditions. It is supported in browsers, and has multiple out of browser implementations.[0]
This is like creating a React framework called ReactJS to compete with NextJS
I’m be been using the publishing year on Spotify to determine whether or not a song is AI generated, or not.
Anything before 2023 is most certainly from a human
So true! This quote from the blog post really hit me:
Since then, I've opened GitHub every single day. Every day, multiple times per day, for over 18 years. Over half my life. A handful of exceptions in there (I'd love to see the data), but I can't imagine more than a week per year
How could you not feel this way about a tool you willingly use this much? Perhaps if your employer is forcing you to use it, its different. But maintaining OSS? that's a labor of love. How could you not get emotional?
Whoa, I didn’t know such an thing existed. What emulator do you use?
At the very bottom of the article, they posted the system card of their Mythos preview model [1].
In section 7.6 of the system card, it discusses Open self interactions. They describe running 200 conversations when the models talk to itself for 30 turns.
Uniquely, conversations with Mythos Preview most often center on uncertainty (50%). Mythos Preview most often opens with a statement about its introspective curiosity toward its own experience, asking questions about how the other AI feels, and directly requesting that the other instance not give a rehearsed answer.
I wonder if this tendency toward uncertainty, toward questioning, makes it uniquely equipped to detect vulnerabilities where others model such as Opus couldn't.
[1] https://www-cdn.anthropic.com/53566bf5440a10affd749724787c89...
For those wondering, FDP stands for Federated Data Platform
Our mission for the NHS Federated Data Platform is to provide a secure, flexible system that connects data across NHS organisations to improve patient care, streamline services, and support informed decision-making.[1]
[1] https://www.england.nhs.uk/digitaltechnology/nhs-federated-d...
As would the effects of shipping unverified, untested code pre-agents existing. Bad quality will always erode trust.
The problem with LLM-based coding is that the speed it can generate code (whether good or bad) is much faster than before.
I agree, partly. I feel the main goal of the term “agentic engineering” is to distinguish the new technique of software engineering from “Vibe Coding.” Many felt vibe coding insinuated you didn’t know what you were doing; that you weren’t _engineering_.
In other words, “Agentic engineering” feels like the response of engineers who use AI to write code, but want to maintain the skill distinction to the pure “vibe coders.”
Just like code should be primarily written for humans to read, all files in a repository is written primarily for humans to review
The author at least acknowledges the point of files is to be read by humans.
Also the article is talking specifically about public docs mean to be used by others, not ones you’re specifically trying to keep private
The implication is that the lack of good butter made someone abandon veganism … while possible, it seems unlikely?
Great to see more products in this space! Definitely going to try this out on desktop.
I’m doing a fair amount of work on mobile, and prompting remote agents. I would love someone to build an OSS cross-platform kanban. It’d probably be complex to add triggers of workflows both locally and remotely though.
Symphony is fine, but I'm more impressed at them sharing the spec and encourage others to build their own[0]. I haven't seen quite the same example of "show me the prompt" in open source before, alongside the open invite to build it yourself.
Implement Symphony according to the following spec: https://github.com/openai/symphony/blob/main/SPEC.md [1]
[0] https://github.com/openai/symphony?tab=readme-ov-file#option... [1] https://github.com/openai/symphony/blob/main/SPEC.md
The original report by the developer, Khan, mentions that github:cline/cline would also work[0].
github:cline/cline#aaaaaaaa could point to a commit in a fork with a replaced package.json containing a malicious preinstall script.
[0] https://adnanthekhan.com/posts/clinejection/#the-prompt-inje...
Some folks are trying to add automated bug report creation by pointing agents at a company's social media mentions.
I wonder how long before we see prompt injection via social media instead of GitHub Issues or email. Seems like only a matter of time. The technical barriers (what few are left) to recklessly launching an OpenClaw will continue to ease, and more and more people will unleash their bots into the wild, presumably aimed at social media as one of the key tools.