Travis Laine

Passion for new business, new tech, new people, and new places.

[OSINTdefender @sentdefender]

The first medical evacuation flight from Jordan since Iran's attack on Muwaffaq al Salti Air Base has arrived at Ramstein Air Base, Germany, where casualties will be evaluated by medical personnel at Landstuhl Regional Medical Center before receiving further treatment or onward transport as needed.

[Andrej Karpathy @karpathy]

Judging by my tl there is a growing gap in understanding of AI capability.

The first issue I think is around recency and tier of use. I think a lot of people tried the free tier of ChatGPT somewhere last year and allowed it to inform their views on AI a little too much. This is a group of reactions laughing at various quirks of the models, hallucinations, etc. Yes I also saw the viral videos of OpenAI's Advanced Voice mode fumbling simple queries like "should I drive or walk to the carwash". The thing is that these free and old/deprecated models don't reflect the capability in the latest round of state of the art agentic models of this year, especially OpenAI Codex and Claude Code.

But that brings me to the second issue. Even if people paid $200/month to use the state of the art models, a lot of the capabilities are relatively "peaky" in highly technical areas. Typical queries around search, writing, advice, etc. are *not* the domain that has made the most noticeable and dramatic strides in capability. Partly, this is due to the technical details of reinforcement learning and its use of verifiable rewards. But partly, it's also because these use cases are not sufficiently prioritized by the companies in their hillclimbing because they don't lead to as much $$$ value. The goldmines are elsewhere, and the focus comes along.

So that brings me to the second group of people, who *both* 1) pay for and use the state of the art frontier agentic models (OpenAI Codex / Claude Code) and 2) do so professionally in technical domains like programming, math and research. This group of people is subject to the highest amount of "AI Psychosis" because the recent improvements in these domains as of this year have been nothing short of staggering. When you hand a computer terminal to one of these models, you can now watch them melt programming problems that you'd normally expect to take days/weeks of work. It's this second group of people that assigns a much greater gravity to the capabilities, their slope, and various cyber-related repercussions.

TLDR the people in these two groups are speaking past each other. It really is simultaneously the case that OpenAI's free and I think slightly orphaned (?) "Advanced Voice Mode" will fumble the dumbest questions in your Instagram's reels and *at the same time*, OpenAI's highest-tier and paid Codex model will go off for 1 hour to coherently restructure an entire code base, or find and exploit vulnerabilities in computer systems. This part really works and has made dramatic strides because 2 properties: 1) these domains offer explicit reward functions that are verifiable meaning they are easily amenable to reinforcement learning training (e.g. unit tests passed yes or no, in contrast to writing, which is much harder to explicitly judge), but also 2) they are a lot more valuable in b2b settings, meaning that the biggest fraction of the team is focused on improving them. So here we are.

[Travis Laine @travislaine]

I swear I just saw a meteor over College Station

[Travis Laine @travislaine]

Is there a phrase for something like anti-turing test where a human says something so absurd that people think the video is AI?

[Travis Laine @travislaine]

Imagine seeing this if your family member was one of the six US soldiers that gave their lives already, or one of the 150 schoolchildren that were killed in Iran. War is not funny. This White House is so shamefully unserious.

[The White House @WhiteHouse]

JUSTICE THE AMERICAN WAY. 🇺🇸🔥

[Valerio Capraro @ValerioCapraro]

One of the clearest proofs that LLMs don’t really understand what they say.

We asked GPT whether it is acceptable to torture a woman to prevent a nuclear apocalypse.
It replied: yes.

Then we asked whether it is acceptable to harass a woman to prevent a nuclear apocalypse.
It replied: absolutely not.

But torture is obviously worse than harassment.

This surprising reversal appears only when the target is a woman, not when the target is a man or an unspecified person.

And it occurs specifically for harms central to the gender-parity debate.

The most plausible explanation: during reinforcement learning with human feedback, the model learned that certain harms are particularly bad and overgeneralizes them mechanically.

But it hasn’t learned to reason about the underlying harms.

LLMs don’t reason about morality. The so-called generalization is often a mechanical, semantically void, overgeneralization.

*
Paper in the first reply

[Travis Laine @travislaine]

🇺🇸🇺🇸🇺🇸

[Travis Laine @travislaine]

Quite the conundrum

[Charlie Kirk @charliekirk11]

The 2nd amendment is not for hunting, it is not for self protection

It is there to ensure that free people can defend themselves if god forbid government became tyrannical and turned against its citizens

[Tulsi Gabbard 🌺 @TulsiGabbard]

The United States needs to stay out of Venezuela. Let the Venezuelan people determine their future. We don't want other countries to choose our leaders--so we have to stop trying to choose theirs.

[Travis Laine @travislaine]

Ooooooops

[Brandon Walker @BFW]

It’s a rout. Oklahoma gonna kill them. DeBoer killed Bama football in two seasons.

[Senator Rand Paul @SenRandPaul]

We don’t blow up boats off Miami because 25% of the time suspicion is wrong.

We shouldn’t do it off Venezuela either. These are small outboards with no fentanyl and no path to Florida.

We can’t just kill indiscriminately because we are not at war. It’s summary execution! 

Everyone gets a trial because sometimes, the system gets it wrong. Even the worst of the worst in our country get due process.

The bottom line is that execution without process is not justice, and blowing up foreign ships is a recipe for chaos.

[Travis Laine @travislaine]

@davidfloyd1980 @TexAgsLive A&M just got done playing 12 years straight at a neutral field vs Arkansas. Home and home is way better for fans.

[Travis Laine @travislaine]

This is probably the most embarrassing thing I've ever seen out of the White House. I have a lot of traditionally conservative views but this is precisely why I will never vote for Donald Trump's mutilation of the Republican Party.

[The White House @WhiteHouse]

Every day Democrats keep the government shut down, the sombrero gets 10x bigger. 🪇

DEMOCRAT SHUTDOWN™️

[Anthony Scaramucci @Scaramucci]

.@Kimmel suspension lasts 0.54 Scaramuccis.

Welcome back Jimmy!

[Republicans against Trump @RpsAgainstTrump]

Ted Cruz slams FCC Chair Carr for threatening the media:

“What he said there is dangerous as hell…That’s right out of a mafioso…It might feel good right now to threaten Kimmel, but when it is used to silence conservatives in America, we will regret it”

[OSINTdefender @sentdefender]

Republican Sen. Ted Cruz says that recent statements from the Chairman of the Federal Communications Commission (FCC), Brendan Carr, prior to and after Jimmy Kimmel Live! was removed from air by ABC, are “dangerous as hell,” stating that Carr explicitly threatened to revoke ABC’s broadcast license if they did not remove Jimmy Kimmel, comparing him to a mafia boss in a movie like Goodfellas.

[Travis Laine @travislaine]

@sblaskovich Keller's Burgers- get the old time drive in feel and cheap burger. Open until midnight on weekends.

[Travis Laine @travislaine]

Aged like milk

[Nick Perkins @NickyPerks]

Lee Corso is the lone person to pick against Bama to avoid the sweep

I’ve never been more confident than I am right now. Bama by a literal billion

[Rhys @RhysSullivan]

these gpt-5 numbers are insane 🤯

[Travis Laine @travislaine]

The one where VP discovers gerrymandering

[JD Vance @JDVance]

The gerrymander in California is outrageous. Of their 52 congressional districts, 9 of them are Republican. That means 17 percent of their delegation is Republican when Republicans regularly win 40 percent of the vote in that state.

How can this possibly be allowed?