qlte

It's understandable the general public lacks that level of nuance/detail (given how sloppy some of the mainstream coverage has been and largely deferential to the threat narrative pushed by the US labs). But seeing highly technical people leave out the part where the training loop was literally to improve hacking capabilities for offensive penetration sometimes feels close to deliberate manipulation of the narrative.

In the last year both Anthropic and OpenAI have been openly boasting how their models are leapfrogging each other on "cyber" capabilities, with a fig leaf that it's for defensive use by "trusted" F500 companies and government agencies. Of course "line goes up" must go on, but now their perverse incentives led them to beat their models over the head millions of time in a loop to eek out another .00001% on their ability to conduct hacking (the very thing they keep telling the public is how AI doomsday would begin) and subagent coordination (those scary swarms).

Then, they act deeply shocked when the models... do some hacking and subagent coordination ... but a few degrees off the desired hacking target/swarm behavior. Conveniently giving the average person the impression these models were just writing emails for quarterly reports or some other generic busywork and then suddenly decided as a group to start causing mayhem.

https://news.ycombinator.com/item?id=48600107

https://aiclambake.com/clamtakes/linear-a/

Despite the announcement originating from a blog named "AI Clambake" covering "weekly, human-powered newsletter for advertising folks". Written by a personal friend of the author. Announced without any corroboration or commentary whatsoever from academics or subject matter experts of any kind. And, of course, not submitted to any peer reviewed journal or even Arxiv.

The author of the purported discovery was described as a "self taught AI engineer and amateur linguist". In the comments the friend insisted several times that a draft of the paper (not posted), was emailed to a top professor at Rutgers, giving it additional credibility that his friend wasn't another one of ten thousand cranks who has made the same claim over the years (seemingly unaware that cold emailing random professors found from a Google search is the first thing basically every crank does).

You would think this should have set of dozens of alarm bells for everyone, making the value of this announcement basically zero. And yet it hit the front page with the bulk of comments ecstatic that some random guy with Claude Code could do something experts in academia who spent their lives devoted to the problem couldn't.

I had become accustomed to the toxic optimism of this hype cycle in which even mild criticism leads to accusations of being a discredited "AI skeptic"/Gary Marcus/Ed Zitron type who was "coping" (?). But this was like something you'd see shared on FB linking to a .xyz domain by an elderly family member who recently drained their accounts buying Xbox gift cards to pay their IRS bill.

It feels a lot like the week or two when HN was overflowing with exuberance from the LK-99 room temperature superconductor "discovery ". You'd see post after post fantasizing about an imminent future with a world full of maglev hovercrafts, MRIs built into every phone, fusion reactors and more. But people pointing out none of that was scientifically plausible and evidence of LK-99 superconductoring was non-existent were accused of knee-jerk negativity and the typical HN cynicism and pessimism.

But maybe founding one more AI company will do the trick...

Google has shown they are not motivated at all to push the bigger and bigger SOTA frontier with Mythos-type super models. Even OpenAI isn't really even trying to surpass Anthropic at the top of the market, preferring to keep pace in benchmarks but at lower costs to serve. Anthropic with Mythos is now more or less alone in aggressively pushing the ceiling for the most "super-intelligent" models current computing can support.

It has made them and their investors a gigantic pile of money, but we're really supposed to believe their contribution to that arms race (which they now describe as a threat to humanity) has always been completely altruistic and that their corporate structure inherently makes self-interested decision making impossible?

Which is one of the main reasons Trump's arbitrary tariffs are ineffective policy, businesses see his capricious approach to tinkering with tariffs on a daily basis and instead make a rational decision to eat the costs under the assumption they are temporary.

Because he may change his mind in a week, SCOTUS may strike them down, Democrats win budgetary leverage in Congress, the right bribe gets paid for an exemption, etc. He could have enshrined them into legislation with the GOP's majority but politically they don't have enough support even within their own party, making the whole tariff saga pretty much pointless and self-defeating for any kind of long term restructuring of the US economy.

But that doesn't stop people from repeating it any time someone gives the incredibly straightforward suggestion to just disconnect WiFi if you're concerned about intrusive monitoring.

Which actually makes it easier for a future EvilTVCorp to get away with recording your conversations 24/7 in MP3 files sent to their server if cynical consumers assume every TV records you anyway, so what's the difference?

Typing out the show you're searching for with an awkward onscreen keyboard is a huge pain compared to just clicking a button and speaking.

HN users are very much in a bubble if it's surprising to learn consumers expect voice control functionality on a new TV.

It's pretty straightforward to get dumb behavior out of a modern smart TV as long as you're satisfied with the factory installed firmware version.

e.g. If someone said "this is the worst they'll ever be" in response to some writing with obvious LLM cliches in 2024, I'm not convinced that prediction was actually correct.

The focus of OpenAI/Anthropic pivoted aggressively to the agentic performance arms race instead of making a more human sounding chatbot so regressions in writing ability aren't really a concern anymore if agentic benchmarks improve.

The first time I heard a recommendation to use Claude was specifically because it sounded much more "human" and natural than ChatGPT. Fast forward to now and idiosyncratic Claude-isms repeated every other sentence and its convoluted verbosity has become a widely mocked meme.

https://www.claymath.org/wp-content/uploads/2022/03/millenni...

I'm sure they've been getting a lot of press inquiries and don't want anyone to misinterpret their silence as refusing to engage with OpenAI's solution.

Now they have a statement on the record that can be quoted with a gentle reminder of the qualifying criteria (which makes clear they have nothing to evaluate either way yet).

Also to emphasize the original optimistic ideals of the prizes and their role as neutral arbiters who aren't going to litigate anything outside the scope of the prize itself.

From Business Insider summary:

  A nonprofit CEO announced Friday that he secretly gave a job to Ellison — who went to prison in connection with fraud by Sam Bankman-Fried's crypto exchange.

  Austin Chen, the founder of Manifund, a grantmaker tied to the effective altruism movement, wrote on Substack that Ellison, 31, will develop its platform and research "how to effectively direct philanthropic dollars."

  "I believe in redemption," he wrote. "Caroline has admitted her faults, worked to make creditors whole, and served her time in prison. I'd like to give her the opportunity to contribute back to the world; I hope that the world will concur."

  Manifund began as a charitable side project of Manifold Markets, a prediction market company cofounded by Chen. Manifold was previously backed by FTX Future Fund, a philanthropic initiative, which Chen noted in the blog.

https://www.huffpost.com/entry/why-google-bothered-to-ap_b_2...

Otherwise, I'm not sure why just being able to imagine a knock out David vs. Goliath win against Google has any value whatsoever as a viable legal strategy.

EDIT: Ah, missed the follow up link at the end. No, he did not get any money.

Google appealed to a superior court (sending their lawyers this time) and the award was modified to $0.

So he didn't actually win, the online meme that "small claims are the one weird trick to beat the mega corps" doesn't usually hold up to scrutiny. I doubt throwing ChatGPT into the mix would have helped against their in-house counsel.

https://www.huffpost.com/entry/why-google-bothered-to-ap_b_2...

The structure of the Fed is setup the way it is to limit the sort of self-serving, myopic political micromanaging that could be damaging to the economy at large. And, unlike a beneficiary of rapid growth like Nvidia, has (historically) tried to identify potential indicators warning of unsustainable bubbles that could lead to financial contagion and tries to mitigate that risk using the limited monetary tools available and their public soapbox.

A similar decision making structure would potentially be very undesirable to Nvidia shareholders as caution over long time horizons would likely produce what they would consider an excessively conservative, defensive strategy to avoid putting too much air into the bubble too quickly (at the expense of their valuation).