boredumb

When I was getting a D&B number I missed their call I wasn't expecting because I wont answer unknowns and my phone blocks half the calls they think are spam anyway, then had to spend the better half of a week answering spam calls before finally getting what I needed.

I do think the frontier models and providers should be aiming to be as insanely accurate and precise for machine interfacing as possible, the rest of the world can build a zillion interfaces into it based on the context that they are actually being used in. That's what they are going to end up doing they just seem to all be trying to build a really great API _for the future_ and a really cool chat bot.

It has worked great but i've spent more time beating LLM output into parseable output than I have reading and appreciating the prose it sends when i'm asking it something about some snippets of code.

I've bet a portion of my current companys platform against this (laboratory operating platform for people doing the biolab work)... uh oh. In all seriousness I think the reality is this is just going to create a lot more work for humans to have to verify and test in lab so I think it will net out in the end as a good thing for me if the barrier for new labs is lowered while the amount of required data and verification lab work is expanded by AI work being published.

Launching just recently and having to think about how i'm marketing things more and more, the hardest parts to really market are the AI related features. Most of the real value AI provides isn't flashy its just eliminating the tedious work that sits between the interesting parts of a job: structuring messy data, extracting information from documents, validating it against references, filling out forms, routing it for approval, and preparing it for distribution. The exciting work was already exciting. AI just removes the friction between point A and point D. That's alot harder to demonstrate than generating an image or writing a paragraph. "Look, here's all the data you no longer have to enter manually" isn't exactly a compelling marketing headline, even if it's where users save the most time.

I sometimes wonder who AI marketing is really for. The people who would benefit from eliminating repetitive work, or other technical people who want to see interesting AI demos.. A lot of product launches end up looking like "Here's a prompt that bootstraps X," when the bigger story is that the prompt replaces an hour of repetitive work that nobody wanted to do in the first place.

Don't work 12 hour days to get milestone X out, because there are dozens more milestones so don't get burnt on trying to get this one out yesterday. It would probably be more like, don't use 200% to get this out and then quit or burn yourself to 0% or a few % in a year when we want you to extend and maintain this stuff.

Cool, but I'm not sure the victims and the public are always thrilled when they see murderers get off by reason of insanity, despite it being a life sentence and essentially a medically induced solitary mental confinement. People were furious about Yates and Bobbitt despite not even living within a 1000 mile radius because they felt it was a miscarriage of justice and are aware that rehabilitating people who drowned their own children is naive at best or yet another grift to siphon public money into their own pet projects and feign moral superiority.

If someone kills a family member and the court gives them 6 years and a parole officer, the remaining family will and has taken justice into their own hands and that has a much higher blast radius and margin of error than executing a guy convicted of the murder in a court of law and sat on death row making appeals for 10-15 years.

If dylan roof was allowed to live his full natural life in jail, there would be race riots in the US by the end of the press conference.

A few years ago I saw a vape with a full display that played a pac man clone aside from the state and settings, and now I have a drawer of random vape screens and components that I swear i'm going to use one of these years.

please, just use html/css and some javascript where applicable.

I thought this was going to be an article about the ESP-32s

just never upset the wrong person that knows they have leverage over you keeping your home.

There is a small sense of irony reading the first 3 sentences before the paywall pops up and stops me from reading it entirely.

as a pedantic aside Americans use the term washed or washing money interchangeably with laundering or laundered money

I've done exactly this with MCP { "name": "unshare_exec", "description": "Run a binary in isolated Linux namespaces using unshare", "inputSchema": { "type": "object", "properties": { "binary": {"type": "string"}, "args": {"type": "array", "items": {"type": "string"}} }, "required": ["binary"], "additionalProperties": false } }

It started as unshare and ended up being a bit of a yakshaving endeavor to make things work but i was able to get some surprisingly good results using gemma3 locally and giving it access to run arbitrary debian based utilities.

keep that energy when you and your family are starving in the dark with no running water.

Who knew the rural town of Remote, Oregon was such a startup hub these days.

Since it says "Translated by Nancy Qanqar" i'd be willing to bet they're training on some audiobooks with a transcript and somewhere in there it consistently has "Translated by Nancy Qanqar" in the transcript where there is dead air in the audiobook.

I went inside and complemented the worker on their pumps being so easy to use, He. Did. Not. Care.

I really don't care about the aesthetics of a string but this is 100% my issue when interacting with UUIDs.

I imagined that the space stations have the equipment to take all measurements and record evidence so you aren't having to return samples to earth except for a neat souvenir.

Climate obsessed people are probably in it for the right reasons but are hysterically annoying from any third party perspective. Keep telling people that cow burps are what is driving bad weather and you can keep being baffled that "deniers" hate any of the solutions you come up with there after. People hate pollution across the board, people hate theoretical models of climate effects based on research divorced from reality that predict weather in 30 years to be a doomsday scenario because of something so obviously mundane while ignoring outsourcing manufacturing to china and india to bus, boat and flight. At this rate we may one day in our lifetime be able to have all the pigs and cattle raised in china and fly them on business class to europe in order to fight the weather via carbon credits.

I've been experimenting with using Tide, sqlx and askama and after getting comfortable, it's even more ergonomic for me than using golang and it's template/sql librarys. Having compile time checks on SQL and templates in and of itself is a reason to migrate. I think people have a lot of issues with the life time scoping but for most applications it simply isn't something you are explicitly dealing with every day in the way that rust is often displayed/feared (and once you fully wrap your head around what it's doing it's as simple as most other language aspects).

> nations that do not tolerate dogs are healthier

Which nations would that be?

I've used chromecast to power all my "dumb" tvs for years and being able to use my laptop or any phone that's on the wifi has been amazing to avoid using a clunky roko or firetv interface. Sad to see one of the most personally useful pieces of google tech ending.

Many moons ago I wanted to do something similar for AI data sets and models over IPFS. I don't know the future for IPFS but I do hope the essence of a p2p data sharing infrastructure becomes more accessible to help individuals tackle some of the issues with large datasets with less hardware on hand.

https://github.com/JakeKalstad/IPFSPytorchDataset https://github.com/JakeKalstad/load_ipfs_pytorch_model

Of course mistral7b is worse than GPT-4, but I can run mistral-7b at home.

> Also I'm curious what would be the legal impacts on it (since USA and EU refers to a model's FLOPs/number of parameters... How do you compute it with sparsity? Do you average?).

How/when did these types of regulations come about? This feels like an insane thing to have to keep in mind while developing.

"Mamba is the world's longest venomous snake with an estimated length of over 150 m"

Had a laugh at that. Really great stuff though, it was nice to have referencing to the arxiv paper so someone like me who generally consumes these things instead of translating them from papers could sort of peak behind the curtains.

Reads like an out of touch journalist with a vendetta against elon musk before I got pay walled.