angry_octet

"Negative results" in the sense of replication failures and trial pre-registration are incredibly important. Lean is in some senses a mathematics response to the maths replication problem -- fields are so specialised, and proof checking so onerous, that many errors will go uncorrected.

Arguably mathematics should have pre-registration, so you can see who has tried what before, rather than just by knowing everyone in your field. As LLMs progress, we might get pre-registration of LLM-aided research. "We plan to spend $10K on Fable tokens to look at conjecture X is algebraic co-homology."

I think the crazy magnitude increased massively. If you subscribe to the horseshoe model, the change in bearing wasn't that large.

If you're not a historian it wouldn't be an egregious error.

She had poise in the moment.

We can agree that he's a massive liar without seeing that as the reason behind everything OpenAI announces.

You would have to go back and do more work though, and if you couldn't fix it you would have to do another defence. But yes, it is a different situation to actually getting a good post doc.

No it's not that they don't check facts, it's that facts don't matter. She found a community where poor scholarship and vibes was celebrated.

I wish I had LLM-built Lean formalisations in university, so much of the math in the slides had errors, and some professors are very bad and ungracious admitting it, while simultaneously rejecting requests for clarifications by saying "the proof is in the slides".

Of course Lean proofs are rarely a good way to understand proofs, but hopefully they can be used to generate more human understandable arguments.

A recording of a car crash: discovering on live radio/podcast that the central tenet of your book is wrong, and amateurishly so.

Naomi Wolf 'death recorded' on BBC[1], skip to 5:51. After this the book was pulped and she had some sort of psychotic break during COVID and allied with ultra-right and COVID denialist loonies.

[1] https://www.bbc.com/news/av/world-us-canada-48639663

Proof fix-ups are quite common, but if it was not possible, then no, not for that topic.

It shouldn't be possible for the US to request this data without the passport being physically present at a US entry facility (border, airport checkin scan, consulate) where fingerprints/photos are actually compared.

Before that point, only passport information tuples should be queryable, i.e., "Is {passport number, name, dob, pob, issue date, expiry} valid?".

We should remember that the current (and likely future) administration is actively hostile to transgender and intersex people, and persecutes US citizens with changed gender identifiers. Discrimination based on ethnicity is only one Overton-lurch away.

Inconceivable!

"Zero Compromise Privacy" vs "Seamless integration with all Kagi services directly in your browser."

Search and translate (and browse?) traces from authentic humans sounds valuable.

No, I am also Silly Bean.

I really dislike this notion that because they had privately discovered a bug, they won't pay a bounty. Better to just sell it.

But now it will match Silly Bean to port3000 and every comment you've made.

Unless you're using short lived SSH CA certificates, SSH key access has only modest security. You would need to buttress it with TOTP 2FA or enforce Yubikey generation/storage of keys, neither of which are at all default.

Even then, it's obvious you are running SSH, and they can fingerprint the OS, and external logging shows which machines are connecting, and hence have said SSH keys. If they have SSH open they become targets; if they come from CGNAT, carriers can leak location via CGNAT behaviour.

In contrast tailscale makes this much harder.

Braided works fine? I have plenty of Ugreen and others from Amazon, what matters is what's inside.

Legitimately, it would work better. I could bulk fax Skills to each of the branch offices on Monday afternoon in preparation for the AI teleconference bridge meeting at 8AM. They can fax back any suggested improvements on Friday.

Because the thermal paper fades you would need freshly faxed Skills every few weeks or the turbo encabulator would misfeed when they were loaded. In fact, I'd have the local steno pool ladies feed in a weekly fax with authorisation codes for each Skill, and without that it just wouldn't load.

Y2K technology for the win.

All the complexity has been in geotechnical issues, pockets of mud, sinkholes, then ultra hard rock, then loose shale, damage in blasting operations, needing to pump liquid nitrogen into the rock to deal with subterranean water flow, soluble limestone etc etc.

One of the scariest industrial deaths you can imagine: cutting into a pocket of shale and the TBM and crew just vanishing into the maw of the Midgard serpent and the whole tunnel filling.

https://www.abc.net.au/news/2024-05-22/tunnel-boring-machine...

Yes Snowy 2.0 just won't have enough water for many of the years in out future.

The submarines will be in port so much of the time, we may as well hook up their nukes to power the grid.

It's to prevent collection of queries from users that are coming from resellers/proxies, for reasons of economy or bypassing region blocks etc. The users are using the stock client and may believe they are using direct OpenAI servers.

I can't think of anything worse than sharing skills via Dropbox. The version management and AIBoM problems that generates is extremely painful. There's no way to track which version LLM is being used or match it against the skill, and people will likely load up too many skills.

You don't have to expose git repos to end users to use git, or some other database, to provision skills.

That is a sensitive messaging campaign, and no doubt true as well. But the effect is that rich people who want to visit on the weekend will book a local hotel stay rather than day trip, and as a bonus they will not have to deal with crowds.

They already have a levy on cruise ships, but it doesn't really increase the amenity for the citizens. If you can bid for places in the cap you have a dual effect, limiting congestion and encouraging longer stays, and selecting for higher value ships.

You could tweak it in many ways, like additional levy that can be recouped by spending on food in local restaurants (vs shipboard all-inclusive meals).

It should be clear that this is about the stress that visiting Indians bring. And their trash.

But it also highlights how you need to restrict access to move up the value chain. Hordes of bus tourists who eat elsewhere or bring take away contribute little economically, you can sell some trinkets. People with a hotel booking are also likely to eat locally.

Venice faces a similar situation with cruise ships and Airbnbs raising the price of housing. They should be capping cruise ship numbers, and a weekend break would be good too.

So disappointed to learn that Americans don't know the game is called Snakes and Ladders, and that their version is all about morality lessons.

Lots of non-work organisations use tools like Slack.

Anyone who has spent time with parrots would realise that they can understand the meaning of speech without knowing what the words mean. Then somehow the meaning of word parts, and then you will find them making new words out of other words. Very clever indeed.

So stochastic parrots could indeed be a good description of LLMs. But I think that she meant it as a diminishing term (against the technology) which is pointless. Probably more of a reaction against SV tech bros than more nuanced interpretations.