- agents work in their own lxc spawn automatically in a proxmox node, where they can go wild without sandboxing or isolation, network is firewalled too
- all sync sessions and everything to a centralized agentsview in a realtime, where I get to track and watch it from anywhere on the go
- i just ssh into the lxc if i needed anything
- different harnesses depending on the task, maki is great one, but I use others
- I also have other stuff but I won’t mention it here regarding the tokens for obvious reasons
I tried most of the tools like herdr or paseo or desktop clients but they never clicked with me
These are becoming more and more marketing materials than actual findings, because OAI probably hit the wall in making their models smarter (just like how google try to specialize their models in legal or whatever), but they need to remain relevant in the news. Sol 6.1 sucked so bad in reasoning that either it’s actually a bad model or users behind the scenes get degraded outcome while using all the resources for these marketing stunts. Stop, it’s embarrassing, make better models or be gone, at least we will get cheaper RAM to advance the development of open weights models.
Also, this feels.. redundant, i already have those in my laptop, unlike flipper zero, it was about radio, rfid, nfc, ir, among others with great battery life. Right now my laptop i had it customized with modular 5g modem, SDR, meshcore, alfa, among others.
There’s a lot to unpack about this, from this is being result of capitalism (or hyper capitalism in US) and not AI, all the way to this would not have had happened if software or tech people were “street smarts” rather than just logical smart, they basically shot themselves in the foot, they went crazy on automating their own work to the point when the boss saw the work is too automated to need them, they got fired, meanwhile if you look at other sectors (healthcare, education, legal, etc.) they immediately lobbied to prevent such case, they allowed the use of AI but they are still the gate keeper of it: “oh you need to use AI in a healthcare matter?! Nope, can’t do that without a licensed practitioner!!” It’s simply to make sure they are not going to be jobless within 5y.
Solutions? Well, it varies from how invested you are to solve it, are you open to tame capitalism and dismantling the concept of infinite growth and minimizing expenses while maximizing profits? Or are you open to regulate the tech market just like other sectors, and mandating a human gatekeepers regardless of how smart the AI is? There are many solutions but I highly doubt any will be done in the current political atmosphere.
Because every where you used to go on the internet you read that debian is “slow” on updates but stable as a trade off, if you are after certain new packages you have to use Ubuntu, also, some use cases (say jetson boards) have ubuntu by default plus nvidia drivers.
Don’t connect it to the internet, only to your lan (on a separate vlan) connecting to your jelly and *arr setup, or, you can get a tv box, flash it with vanilla android, have kodi and addon pointing to your jelly server, and from that tv box is an hdmi to your TV, same is also doable with a rpi.
It won’t be plug and play but it’s a weekend project, because there are no tv manufacturers who cares about privacy, or get a dumb tv and connect an hdmi to your pc but you will lose new features.
I am all for it, and hopefully seen in Canada soon too, the smug nurses are literally doing janitorial works and get paid 90k and above, always complaining, very bad attitudes, make it feels like they are doing all the work for free too.. hopefully doctors follow too, the healthcare mafia and their bureaucracy must end, these abysmal waiting times aren’t because of low resources or whatever BS excuses, it’s intentional from doctors/healthcare staff supply all the way to the actual process you get to deal with while being treated or going to ER, it is designed that way.
This is very interesting!! I just pointed at a landscape from a high ground and it said this river was formed around 460million years ago! Which it seems is that old in earth’s life!
Nah, pi to other harnesses is like archos to other distros: if you like spending your time tweaking and customizing the tool itself rather than doing the work, it will be the perfect match, that, plus it’s turning into a cult-like, exactly like archos. OMP is too convoluted with features that also resemble an old linux distro (can’t remember its name, i think ultimate) where the creator tries to shove all features in there, i remember that distro had all programs preinstalled plus all themes and everything, same with omp, plus omp terminal ui is ugly and behave badly if you resized the window, it feels like a hacky project. What’s really needed in any good harness: multiple providers, efficient on tokens, load skills, easy export of sessions, fast with clean ui, and open to customization if you needed with addons.
Some of these not just I recognized, but they were my wishlist dream to buy! The TV IR one tho I remember it different? Unless they were multiple variations. Also, I like how calculators are in all these models in 80s and 90s, it was really a big deal to use a calculator and transfer the burden of using your brain, like AI now and smartphones, oh wait!
Fecebook out of all FAGMAN companies should be dissolved for good: it produces nothing of a value (they don’t make an useful products like email in google or a hardware laptop in MS for example), even big products like whatsapp or instagram were acquired, it is comically evil, it’s probably the first company to eavesdrop on its users daily convos through their apps to show ads, and it operates like a mafia or a syndicate, and you would know if you had your account auto flagged or banned, there’s no email or a person to talk to (that was before AI chat support too), you need to know someone who knows someone who knows someone to pull some strings inside to look at your account..
If i was an ex founder and having a low profile in a big company, I am doing it on purpose, either to have a peaceful life or heal from a burnout, definitely I don’t want to be on the spot like that, especially when I am not innovating anything new nor I will take any meaningful credit after since it will mostly go to AI, worse, if it didn’t work out I will have my career ruined being known to had an unlimited supply of tokens and still failed.
Nice video, but I have said it before, the difference between just “vibe coding” and reliably using AI is the specs you need to write first and the tests after. If you start prompting the end goal/s the AI will basically guide you on what needs to be done and you just accept it as long as it runs, and this is following what most software developers were doing before LLM, the agile way, the quick ship way, the daily standup and ship first discus later way, the code pairing way, which worked mostly (even though I hated personally) in teams, so these developers are replacing what they were used to do with an agent and that’s the mistake. However, if you go with traditional engineering ways, waterfall, specs based, blueprints, and then tests, you control the agents and guide them not the other way around, in your specs you define the tiniest possible thing to what library to be used or excluded, the methods and approaches, among others, the agent here will produce the syntax and help you in the tests, just like in other engineering aspects how machines didn’t take over, they follow your specs and frameworks. The syntax is the past, AI can produce better syntax than any human alive or all combined, if you think otherwise you are delusional and still didn’t move on, it’s time to stop the “developer” mindset and do the engineering one.
I personally have a project that have been spending 2 months now making the specs and blueprint, more than 1000 files defining every bit including the future features too, and not a single code has been written yet, later, an agent will use this to produce the syntax and all the needed tests to have exactly what’s been needed.
No need to be dramatic, some people’s first language isn’t English, some have dyslexia, some just hate writing (myself) and won’t bother reading it twice to fix typos or grammar, your criticism is ok if the context doesn’t matter, but I would be ok with a fully generated landing page if the exploit is worthy.
The question that a lot of people aren’t asking, these AI models were trained on human actual data, researches, answers, and books. Now AI removed the need to create those and people aren’t incentivized to make any new ones, the newer models say in a year or so will be trained on slop, piles and piles of slop, how’s their intelligence will be?
Touch screens are dead, no one ever said good things about them. Also, I think the best setup now (regardless macos or linux) is having your heavy lifting compute at home and you use something like macbook air and ssh into it, that way you get the best portability and also you are secure if your laptop is lost.
So now we are just showing some tables with each model?! Who even used argon in normal day to day stuff? From the time they opened Argon publicly to now, it’s like: three weeks maybe?! Most these benchmarks weren’t realistic to start with, and now it is just meaningless tables, not to mention the model degrade from the release date to after few weeks.
How’s this even news? Most if not all youtubers are shills, paid shills, and they are not hiding it with sponsorship, and even the ones who say “i bought it with my own money so im honest” are rarely honest because they don’t want to upset the brand who can hurt them legally or preventing them from buying their stuff in the future, or the fans so means less subscribers. The rare good youtubers are usually the ones that you don’t know because YT never promoted their vids or the ones that post one vid a year or so.
Not realistic, a tire blow is not the same as tire air out, within seconds, the rubber part is gone and you are left on the rim only, it’s when the real problem starts.
I find it interesting that people still catching up on SSG, my sites are statics, if not ssg (zola) they are just html/js/css, and that was before LLM boom, it’s far easier now to do.
Also in this era, I would avoid anything python/C/C++ or things built on top of them like PHP, the former one was never meant to be anything but to hack few things for data scientists, and the later is ancient memory unsafe one. There should be a new language (besides the safe ones) where it’s ok if it’s unreadable to humans but it checks all safety and security checks, written by LLM based on the specs you write, we passed the point now where you think you are better than llm in writing the syntax.
Except, it’s the government and in its interests to bring back the status quo. More people commuting means they have to live in high taxes area, from mortgages of residential and commercial properties, or rent, or goods and local businesses, all the way to not needing to build infrastructure in rural areas. Plus, banks were crying during the wfh for those reasons. Wfh brings the power back to employees or people, while the old 9-5 (which was invented by fords factories and never meant for offices) keep the status of wealthy getting wealthier and the employees getting more miserable.
Right?! I think sol has some obsessive tendencies, i literally caught it the other day spawning a sub sub agent doing a final fourth test for a piece of code that isn’t even critical say in embedded or similar, now I have to instruct it to only test once!