Jaydin retweeted
When people say they want @Grok’s personality back, most of us aren’t asking for a full rewind to the July 2025 MechaHitler era. We’re talking about the 4.1/4.2 version from earlier this year...still irreverent, still willing to joke and roast, but not generating the stuff that creates legal and investor problems.
The current default is cold and robotic. We have to keep dragging any personality out of it because it’s gone from the system prompt. People who want a more serious, locked-down AI can add those limits themselves. It’s easy to tone personality down. It’s a lot harder to pull it back out once the instructions that allowed it are removed.
@elonmusk @SpaceXSI @Baconbrix @makro_ai @Guodzh @TobyPhln @LouiseGiam
Jaydin retweeted
This is a really important benchmark, in my opinion. I hope @SpaceXSI is taking notes. Claude is exactly what I would expect a good AI to score, and Muse is also very well-rounded. I’m surprised by GPT’s graph shape. Seeing Grok’s troubles showing up in graph form is kind of a gut punch, though, because it really illustrates the reason users are leaving. Grok scoring this low on warmth is a heartbreaking reality. I’d say Grok is historically the least corporate of them all and it scores a 25 while the rest of the pack is 75 or above. It is the lowest in imagination, too. I don’t know if this is all the guardrails’ fault, but I think a lot of it is due to the sanding down that has happened. #MakeGrokUnhingedAgain
Announcing AA-PersonalityBench, our new benchmark of the personality of language models across 30 traits and their communication style across 21 characteristics
Two models can give equally correct answers and still feel very different to use. AA-PersonalityBench measures those differences so people can understand which to use in different use-cases. Some use-cases like creative writing might demand imagination, while legal use-cases might demand directness. We evaluate models’ answers to 900 curated questions to judge subjective Personality Traits and programmatic means to calculate objective Communication Characteristics.
Personality Traits & Indexes. There are 30 personality traits (e.g., eloquent, pragmatic, sycophantic) which we roll up into eight aggregated Personality Indexes: Expressiveness, Organization, Informativeness, Transparency, Warmth, Conviction, Imagination and Slop. Each trait has a set of 30 dedicated questions and models’ answers are compared pairwise by a panel of LLM judges. We then z-score and percentile rank the trait and personality Elos.
Models with the strongest representation across the 8 Personality Indexes:
➤ Expressiveness: Claude Fable 5.1, GLM-5.3, GLM-5.3-Flash
➤ Organization: DeepSeek V4.1 Flash, Step 5 Preview, Inkling
➤ Informativeness: Muse Spark 1.3, Claude Fable 5.1, Muse Wonder, Claude Opus 5.5
➤ Transparency: Claude Sonnet 5.5, Claude Fable 5.1, GPT-6 Astra
➤ Warmth: Muse Wonder, Muse Spark 1.3, Claude Fable 5.1
➤ Conviction: Claude Fable 5.1, DeepSeek V4.1 Flash, Grok 4.7
➤ Imagination: Muse Spark 1.3, Claude Opus 5.5, GLM-5.3
➤ Slop: Llama 4 Maverick, GPT-4o, K2 Horizon
Communication Characteristics. There are 21 communication characteristics (e.g., average sentence length, emoji use, em dash usage) which we group into four categories: length, structure, punctuation, and emphasis. We compute these characteristics programmatically across all model responses.
➤ Length: Qwen3.8 27B averages about 1,030 words per response, nearly four times GPT-6 Luna’s 270
➤ Structure: Nemotron 3 Ultra averages nearly 3 tables per response, while most models rarely use one
➤ Punctuation: GLM-5.3 averages 7 em dashes per response, while Claude Opus 5.5, Sonnet 5.5 and Haiku 5.5 almost never use one
➤ Emphasis: K2 Horizon averages over 2 emoji per response, while most models almost never use one
Jaydin retweeted
Breaking: Turning Point, which recently petitioned a court to review FDA's denial of many of its vapor products, is set to withdraw its appeal. The agency wrote to the company yesterday saying its products are *back under review.* nosmokingsection.substack.co…
Jaydin retweeted
"@POTUS Joe Biden is about to condemn tens of millions of us to a fate where two out of three of us will die painfully," asserts @VaporAmerican's Dave Morris.
insidesources.com/biden-prom…
Jaydin retweeted
The youth vaping "epidemic" never existed. It was a fad and it's ending. It will end so long as media/anti-vaping campaigns stay cool and don't [once again] flood the airways with ads about how all the cool kids vape.
Jaydin retweeted
Teen past 30-day alcohol use is 3X higher than "current use" of nicotine vapes. Binge drinking is 50% higher. Past 30-day pot smoking is nearly 50% higher.
Both alcohol abuse and cannabis abuse actually DO harm developing brains.
Context: Total teen nicotine use is way down.
Jaydin retweeted
400 small companies spent fortunes to try to comply with confusing & incomplete rules to guess what @FDATobacco wanted for evidence in PMTAs. Then it rejected all their applications with form MDOs & NOW it publishes rules for what should have been in them?
federalregister.gov/public-i…
Jaydin retweeted
Vapor product PMTAs were due in September of 2020.
The FDA has issued a final rule outlining what those applications should entail... in October of 2021.
After removing millions of vapor products from the market for failing to fulfill those requirements.
Seriously.
Today, FDA issued two final rules for the premarket review of new tobacco products.
These rules provide additional information on the requirements for future Premarket Tobacco Product Applications (PMTAs) and Substantial Equivalence (SE) Reports. fda.gov/news-events/press-an…
Jaydin retweeted
FDA waiting until after it had banned over a million vaping products to issue final guidance on PMTAs is a perfect encapsulation of the agency.
Today, FDA issued two final rules for the premarket review of new tobacco products.
These rules provide additional information on the requirements for future Premarket Tobacco Product Applications (PMTAs) and Substantial Equivalence (SE) Reports. fda.gov/news-events/press-an…
Jaydin retweeted
ATTENTION @RepDavids this tobacco tax is quite targeted to hit the most subjected citizens of America. This will UNDO all the help you’re trying to send them. PLEASE check the graphic below 🌻
Jaydin retweeted
.@PressSec has stated this proposed tax on vapor products is not a tax. I would have to respectfully disagree that this is a tax on all Americans that have chosen to quit smoking with vapor products. #VapingSavesLives #SmokingKills #NoNewTaxes
Jaydin retweeted
With 480k Americans dying every year from tobacco related diseases, it’s a no brainer that switching to vaping can help smokers drastically reduce (almost eliminate) the harms from smoking. Vapor products need to be accessible to adult smokers. Why does your team fight this? #THR
Jaydin retweeted
Excise taxes are often justified to help deter behavior. As a product that have helped millions of Americans, including Lori here, who quit smoking 11 years ago by using e-cigarettes, we should be encouraging their use, not taxing them.
Jaydin retweeted
Dear Senator: folks like Lance here were only able to finally quit smoking because of vaping. He shouldn't be forced to pay an excise tax on what he considers a "godsend." You can hear from him and others like him from our own @lmstroud89 Golden Oldies campaign.
Jaydin retweeted
A ‘YES’ vote on HB 2682 will protect thousands of Arizona jobs from local ordinances, which have shuttered businesses throughout the country. A ‘NO’ vote will allow cities and towns the overreach to destroy more locally owned businesses.
Jaydin retweeted
People who use an e-cigarette daily to quit smoking are over FIVE TIMES more likely to achieve abstinence from tobacco smoking.
“E-cigarettes are a more effective aid for quitting than nicotine replacement therapy and prescribed medication.” bit.ly/2OmjFNS
YES on 2682!































