The Mosaic browser (1993) had full text history search.
We then got a bookmark system that was every bit as terrible as a web directory.
It stayed that way. The delicious search revenue made organizing websites uninteresting. That obscure website you enjoyed a decade ago but don't even remember, they had lots of traffic like you. No point updating or keeping it online. You can't have rss in Firefox but here is a Facebook like button in your address bar in stead.
I've tried to maintain the bookmark menu but I rarely use it since everything is dead. Why aren't browsers storing a text version of the bookmark? Did people in 1993 have more resources than us? Should I be afraid it grows to a few GB over the decades?
A good few dead websites have a backup some place but there is no automation to find it. If you had a string of text from a page you might be able to search for it. If the page found is highly similar we might automate the process to have alternative location for bookmarks with a nice warning dialog.
The Office of Weights and Measures exists because, long before any of us were born, in 1836, companies were up to shady shit and consumers were paying for inconsistent products. I.E. Being scammed.
AI companies should be subject to the OWM like any other company that sells a product that varies in weight. Perhaps when a sane administration is re-elected; one that can read history books and comprehend why our regulations exist in the first place. Or have even a semblance of respect for its citizenry.
I will never touch this garbage and would rather use the disk space I purchased to be used for my purposes, not the industry's pointless fad bubble endeavors.
Bookmarks died because most humans are allergic to having to do bookkeeping and cleanup tasks that are incidental to what they really want to do.
We see that over and over again, in a bunch of non-tech fields. My kids never want to clean up their toys; they will pull out new ones, but it is a struggle to get them into the habit of putting the old ones away first. My parents used to keep a notebook with every gas fill-up they made; nobody born after about 1990 does that anymore. We were taught how to balance a checkbook in elementary school; basically nobody does that anymore, we put everything on autopay and if you are diligent you check a statement or import it into Quicken once a month (most people don't even do that, they have no idea what they are spending and predictably, usually no money left over). GMail succeeded because instead of putting your mail in folders, you just leave it in one lump with Google and rely on full-text search. A lot of Zoomer computer users don't even know what files and folders are, they just use apps, which take them straight to what they want to do and don't offer things like possession of your own data.
Back in the (~early 2010s) days when Google still allowed internal innovation, there were recurrent demos produced by engineers of full-text search over your web history, though of course it was your web history as stored by Google and none of this was local. It never became a product, largely because users are too lazy to go to a separate search product just for your history, or because they're too lazy to check a separate box saying "Search my history". Instead I think web history became a ranking input to general search and it would mix in results that you frequently visited to the general results, which honestly I think was a more useful approach.
I really don't think it's productive for internet forums to constantly be criticizing language choice when the meaning is clear. Better to respond to the substance of the issue than word choice.
Edit: In response to the initial replies. To me it clearly means "releasing frontier models at any pace less than as fast as possible". It implies relative restraint compared to the previous state and without stating the degree of restraint.
I cut out Social Media earlier this year, and it's been one of the best decisions I've made. Part of why I started was remembering that we didn't have doom-scrolling apps before smartphones. The change has been tremendous since I've been trying to go back to more "intentional" media. When I find myself distracted and wanting to pick up my phone, I realize there isn't really anything on it for me to do and am forced to just put it away. So I've started getting into the habit of keeping a book on me wherever I go. I've read 4 books this year!
There were sellers buying the smallest RAM units, changing the RAM chips, and then reselling them as the higher RAM SKUs. The sellers who do this don’t always care to use good RAM chips and may even use QA reject parts. When the unit doesn’t work correctly, the anger and RMA requests are directed back at the Raspberry Pi foundation.
For us dell has become that. Rep changes all the time, what appears to be random pricing and long unclear delivery dates. Instead I can for example go to Thomas Kren and configure my serves online for a fraction of the cost. They have what they list in stock, delivered a few weeks later. Everything is clear and hasle free.
At this point I have the feeling dell just charges you what they think they can get out of you. You push back and the price drops, I don't want to negotiate.
No, OpenAI did not solve the "wrong" Navier-Stokes problem. OpenAI did not solve the hardest version of the problem (unforced blow-up), but did give a solution to the Clay Millennium Prize Problem as written and understood, choosing the explicitly allowed forced option.
SciAm writes "in a sense, the LLM found and exploited a loophole in the framing of the question". This is pure sensationalism. Choosing option (C) (out of an explicit list of four options) is neither a "loophole" nor something "found by the LLM"; everyone involved knew this was the option they were pursuing.
With the grumbling out the way, there is some actual scientific content to the article: there's a strong argument that OpenAI's method will not extend to the unforced case, leaving our understanding of NS incomplete. This negative result is itself new and interesting (and predicated entirely on the solution found by OpenAI)!
I... don't know how I feel about this. I mean, the concept, sure, good idea. Is it reliable? Perhaps. But... I get a "buy merch" popup the moment I open the page, and there seems to be a pro option for what seems to be a very much LLM-aided app? ...IDK, rubs off a bit wrong IMO. Just my personal opinion.
(ass. sports.txt politics.txt and business.txt are text docs pertaining from the sports, politics and business domains, respectively, and have equal size)
The test file belongs to the topic with the smallest size *.gz file.
Witten's group at Waikato uni were perhaps the first to work on this.
Also check out the Hutter prize if you are interested in this.
Every major AI shop has a ton of in-house classifiers already, big, small, generalist, specialized. Some are used in inference pipelines (e.g. safeguards), some are used in data preparation, training, analysis and investigation, research, various one-off and intermediate tasks etc. Offering them on a public API doesn't always make business sense. I don't see much substance to this buzz, looks like people that are new to all this are discovering that classifiers exist, they are more efficient at classification, and many tasks commonly done with generative models are classification in disguise. Which is not bad at all, a fresh look at their use is great to have.
There’s a scene in Battlestar Galactica (2004) where someone asks Captain Adama why the Galactica doesn’t have networked computers. So the cylons can’t hack the ship…
PS5 emulation has gone from barely working to running Dark Souls at 10+ FPS with virtually no graphical glitches… in 6 weeks. If you’re not familiar with normal emulator development time, this is… quite extraordinary. There have been insane progress on decompilations and many other things in the emulator space.
Everyone’s trying to figure out how to convert this speed to product features at scale, but enterprises are like container ships. Lots of might but slow to turn. The littler companies can actually take advantage of this and produce higher quality products at much faster speed. I think you’re expecting too much in the short term and too little in the long term. AI-native companies are gonna eat everyone’s lunch, once they figure out how to actually do it reliably.
I have to be honest. While this is obviously a smart and useful idea, it misses one of the core features of Jev: its confidence scores. Partial confidence could easily be mapped to fractional spaces, using unicode characters like U+2009: THIN SPACE. As it stands, this package is not harnessing the full power of Jev.
It's not just a scene; it's the whole premise of the setting. It's why the Galactica survived and the newer ships did not. It's why the new Vipers got wiped out and they had to pull the old ones out of mothballs.
In the pilot, the Galactica was literally being turned into a museum, and that's why they lived.
I might turn this into a blogpost if folks are interested, but my god there is so much clever info in that dashboard.
Here is one really neat bit:
A cutting edge training idea (for agents, it's been used elsewhere for ages) is on-policy RL, basically, it's not enough to say "here is an end to end agentic sequence (including tool calls etc.) that is perfect" you want to say "here is a sequence you might actually have generated that turns out to be correct".
Basically, it's more training efficient to improve models with small tweaks to do more of the right thing they are already doing sometimes than from some perfect oracular "this is the way" answer.
(if you've ever tried to teach humans new skills, you’ve probably noticed this too!)
When you do that, you care about how far the model you are updating (improving) has deviated from the one being used to generate rollouts (agentic rollouts for hard problems can take hours with lots of tool calls, so you can't keep redeploying every slight improvement).
Lo and behold, the dashboard literally has:
partial/avg_staleness (likely the measure of how many micro iterations the "generate answers" model is behind the "improving based on the occasional right answer" model)
train_infer_diff/new_infer/kl (a more direct KL divergence based way of measuring how differently the two models generate tokens)
How cool is that?!
And don't get me started on the clever ideas hiding behind dynsam/avg@n ...
I would never have figured out the toggle for writing tools is in `Settings -> Screen Time -> Content & Privacy Restrictions -> Siri -> Writing Assistance`
If you want to own your computer, linux is where it's at. Windows is rusty iron chains and MacOS's are gold-plated silver pretending to be gold, but they're still chains.
>A man stranded in the bush in northern Saskatchewan was rescued last week after chopping down four power poles — knocking out electricity to surrounding communities. [...]
>But he had an axe and he knew SaskPower would have to check the downed line, so he went to work.
> "The prediction was that software engineers would lose their jobs first, and the recent trend went the other way: demand for software engineers is higher than ever"
Hard to take anything the author says seriously making nonsensical claims like this. The software engineering job market has been getting worse every year since 2022 by virtually every metric. This is especially true at the entry and mid level. For example, computer engineering and computer science majors now have the #2 and #4 highest unemployment rates amongst recent graduates [1]
Students are smart to be cautious about the future, and it's annoying that adults with no skin in the game so flippantly dismiss these concerns without any data to back it up.
So Uber ToS requires you to accept arbitration, then, when they are found responsible for damages, they still don’t want to pay. Seems pretty shitty for the consumer.
As is common in political shenanigans, the draft bill has a euphemistic label ('algorithmic opt-out') - what it actually does is allow unspecified bureaucrats such as the Minister for Communications to designate any individual piece of content as 'harmful' and have it banished from the internet.
As an Australian who just wants relatively free access to information (and the ability to freely/easily discuss it), it's frustrating.
And it's not the first time the Australian government has done this; last year they banned U16 from social media, several weeks later everyone was alerted to Bondi by... social media.. so they could get the heck out of there, but U16... they were left clueless and unable to get updates (news sites ran 30+ minutes behind real-time events and official sites like NSW Police took 70+ minutes to even announce there was a shooter).
We then got a bookmark system that was every bit as terrible as a web directory.
It stayed that way. The delicious search revenue made organizing websites uninteresting. That obscure website you enjoyed a decade ago but don't even remember, they had lots of traffic like you. No point updating or keeping it online. You can't have rss in Firefox but here is a Facebook like button in your address bar in stead.
I've tried to maintain the bookmark menu but I rarely use it since everything is dead. Why aren't browsers storing a text version of the bookmark? Did people in 1993 have more resources than us? Should I be afraid it grows to a few GB over the decades?
A good few dead websites have a backup some place but there is no automation to find it. If you had a string of text from a page you might be able to search for it. If the page found is highly similar we might automate the process to have alternative location for bookmarks with a nice warning dialog.