working on my own AI harness https://github.com/syrull/pluto tldr I dont trust anything nowadays, I wanted to know what requests I am doing to where and basically implement features that I use daily without the clutter of the other harnesses, still very raw but I am using it exclusively. Not a product or anything just something for myself.
I feel like the topic just scratches the surface; there are many more things that are way more interesting to talk about here.
For example: air density actually matters. Professional players sometimes struggle to make shots in different parts of the world. The balls are not the same either. Comparing DHS with Nittaku, for example, they have different properties and a different bounce. In my part of the world, DHS is the preferred ball, compared to the US, where, at least where I’ve been, they prefer Nittaku balls.
Now about the content:
The Bounce: This could have shown all three balls at the same time, instead of showing them one after another. That would make the effect clearer.
The Answer: This feels a bit odd, and I think something in the calculation is wrong because the usual push always goes over the table on backspin.
Overall, I get that this is a very simple representation of what table tennis is, but it gives me “I did it in 5 minutes with AI” vibes.
I had an interview with a CTO back like 2 years ago, I already passed the 3-4 interviews 2 of which were technical and I am on this cultural fit interview with the CTO. The guy enters the call everything seems normal and he begins with his questions which were absolutely insane. First he asked me how did I improve in my work, since I work in cyber I explained to him that I am following latest trends, and getting security certificates like OffSec and other hands on stuff, even CTFs. He is not satisfied he asked me about specifics and to dive deep, okay.. I've explained to him how OSCP, OSWE and OSEP worked and possibly even shared some of the exam ranges scenarios, again he wasn't satisfied and asked for specifics I am already baffled and I was thinking "Does he want me to share the CTF flags or what". Anyway after back and forth with me saying that I don't understand what he wants me to say, he moves forward with "Now what about outside working hours". I sighed and explained that this is taking at least 90% of my free time as well, but he kept asking that he does not care about work and wants to know what I do outside work. Okay I play table tennis quite often and I train it, HE IS NOT SATISFIED (mind you he asked for specifics as well) and in the end he sighed and just said "so you are not improving outside work". I am on the verge of clicking the end meeting button but I kept going. Next question was, have I done any proactive work during work hours, I explained what I did, what tools I've written for the developers and he again asked for SPECIFICS, I said that at this point I am starting to worry about my NDA, and explained that I won't be sharing what I did specifically. He was frustrated and started explaining me that he does not know my clients or whatever or anything about me so it won't matter, which is absolutely laughable but I refused to share anything. Next thing he said is that I am bombing the interview and he has 0 value from this conversation, and I was like okay sure I didn't know that I should bring value into your company/you without even starting. He was now really frustrated started to use words like "fucking" and so on, explained to me how the interns in the company were having a better answers and how his mom would do a better job at this. I was like "???? lmao". Now telling you this I am not sure why I did not quit the meeting but here we go the next question, which I do not remember exactly but I distinctly remember that he said "that's a perfect answer". And since this is becoming long enough I would spare the rest of questions which were absurd once again. Needless to say that I wasn't chosen nor I would have chose them even if this was somehow a pass.
Exactly my thoughts, I believe that the 10x engineers are basically living in a bubble and completely oblivious to the average developer/engineer (I consider myself average). Yes the LLMs cannot bring the level of sophistication that geohot would have but they totally could satisfy the needs of an average developer and their average job. More than 95% engineers are not having the problems that geohot is solving and most of our works is straightforward that any LLM can do, thus enabling us to do more of the same and possibly focus on slightly abstract problems if we have the time. Someone said couple of months ago that manual coding will be a privilege and I see that now.
inb4 "I got prompt injected and they stole my stuff". Now real talk, there are some viable usages of codex here but nothing novel its the same "old": "MEMORY,VAULT,BG TASKS" that everyone is doing.
And about voice mode, I thought it was a good idea but I seriously don't know how you guys use it, my thoughts whenever I use voice are "aaaaaaaaahhhhhh, uhmmm" and then cancel it so that I can type and organize my thoughts. I don't really think those "brain dumps" are useful when you are thinking out loud like "We should really do X oh wait but actually Y is in the way and we have to take into consideration Z, but wait Y was actually done" and so on, and it turns out that your assumptions are wrong, it becomes a mess. I am in favor of the LLM to work with facts and always verify it. To me this post is basically selling Codex app and that's it, nothing new inside.
I feel that few will have the privilege to have the time to write code by hand. And let's actually see what we are actually writing, most of the time for me its nothing novel, nothing fancy its the same old create a backend for X, fix some simple bug and stuff that are trivial for a mid-senior programmer. The harder tasks are mostly (again for us) architectural decision over the code and I am even thinking of how we can develop a system where LLM wouldn't derail on feature implementations. Anyway, what I am trying to say is that you writing code by hand may be okay for now but in the future I believe the shareholders and whoever is on top of you will want you to deliver features/fixing bugs FASTER with the help of LLMs and if you can't deliver that you will under perform. So in the end it's not what we want but what the shareholder wants. Of course if you aren't drained by this you can write code by hand in your free time. I don't want to sounds like a doomer but I believe this will be very much a reality sometime soon.
That's pretty good, I've achieved pretty much the same thing using the vercel's agent-browser, but I've tried playwright and it worked easily as good. Its good for scraping, automating stuff in the browser.
Yea, the thing is that currently this system works only for my country (Bulgaria) and to expand this further and to be effective deeply requires local integrations, but then again there are so many problems to be dealt with.
I am building a platform for helping stray dogs/cats pro actively. The idea is whenever you see an animal in need in an unfamiliar area, you can use it to help the animal get into the right hands. Meaning that you share your GPS and it automatically sends you the nearest clinics, relevant government institution, phone numbers of people that are monitoring for such things and local facebook groups that you can post on. I am also exploring a ways to post automatically in a facebook group whenever a signal is received. There are a lot of challenges however with the platform as it needs to be mobile-first and easy to use for the people to actually start using it.
I am actively using ohmypi harness which is based on pi-mono which I believe is within OpenClaw, I don't personally use OpenClaw but I suspect that I will be affected. The reason that I use ohmypi is because I can extend it and put guardrails specific for our company and myself (those are different from SKILLs and more sophisticated than the hooks) + I like the ability to start "tasks" with faster models like gpt5.4-mini for certain tasks and overall have the multi-model capabilities, now all of this seems impossible. I have the $20 sub from OpenAI and it seems that the usage is similar to the $100 plan by Anthropic, I am extensively using GPT5.4 to review and sometimes code along with Opus, right now it seems to me that OpenAI is winning, I can just go with the $200 unlimited usage by OpenAI and use 5.4/5.4mini for everything. On top of that the Chinese models are really capable at the moment, I've tried StepFun and it's really good. Seems to me that Anthropic is sabotaging themselves with those moves. But it is what it is, the cycle of model switching has begun again, I strongly believe that in 2-3 months they will revert that and we will switch models again. :D
Yes, but as a producer I would like to have more simplistic generations such as "Generate me 15 variations of a kick that sounds like X", I think stuff like this would be much more useful.
I've tested it just now, very Opus-like experience. The speed is also there so far I think I even like the response of GPT5.4 better than Opus (although very close) I might not distinguish them just yet.
I tried several use cases:
- Code Explanation: Did far much better than Opus, considered and judged his decision on a previous spec that I made, all valid points so I am impressed. TBF if I spawned another Opus as a reviewer I might got similar results.
- Workflow Running: Really similar to Opus again, no objections it followed and read Skills/Tools as it should be (although mine are optimized for Claude)
- Coding: I gave it a straightforward task to wrap an API calls to an SDK and to my surprise it did 'identical' job with Opus, literally the same code, I don't know what the odds are to this but again very good solution and it adhered our rules of implementing such code.
Overall I am impressed and excited to see a rival to Opus and all of this is literally pushing everyone to get better and better models which is always good for us.
No not really, I rewrote that part since it gives the reader the wrong vibe. The RCE is quite unlikely (although possible), I believe however that people at OpenAI should care for such "P5 vulnerabilities" since something minor as this could be chained into something else later on.
Security Engineer / Software Engineer with about 10 years of experience. I've designed a lot of open source tools and worked on major EDR solution. I've worked in big enterprises and startups through my career. I hold OSWE,OSCP and CRTO security certificates. I know my way around in almost every major programming language and framework. I am very proactive person and I am always giving my best to improve.
I am interested in building the next generation of agents and I am looking for interesting AI work. The security domain would be a plus but not mandatory.
I have the exact same problem and I am using the same tools, I was thinking of how to solve this, and the best thing so far is to use Claude with the Notion MCP.
My workflow would be to open claude's app and just use the voice tool to say something like "create me a note about blabla in my personal notes", this with a custom project where the system prompt would be about note taking and how to organize stuff pretty much solves this.