The data science is where the real value comes in. 10x flags changed this release - which one caused the improved CTR? Booleans as a service need to address this, and there are benefits to having your boolean service live next to your other services
I think it's interesting that folks are suddenly taking issue with "anthropomorphizing" language used in AI as if we haven't been doing this since the earliest days of computing (see "memory", "child", "parent", etc). It helps folks understand things at the correct level without needing domain knowledge
This kind of continuous fine tuning is being actively researched very heavily, but you might not see much about it at a conference because the labs don’t have enough to show their cards yet.
> Taxation without Representation was literally the driving force for the creation of America itself
The issue of taxation without representation had far more to do with the founders’ status as Englishmen and British subjects than their status as taxpayers. Paying taxes by itself was not a sufficient qualification for political representation. Felons, minors, and women were also required to pay taxes in the 1770s, despite not being able to vote. Immigrants who believe that the taxes they pay entitle them to this representation have bought into a falsified version of American history that was popularized during the Civil Rights Era.
United States Congress, “An act to establish an uniform Rule of Naturalization,” March 26, 1790:
> Be it enacted by the Senate and House of Representatives of the United States of America, in Congress assembled, That any Alien being a free white person, who shall have resided within the limits and under the jurisdiction of the United States for the term of two years, may be admitted to become a citizen thereof on application to any common law Court of record in any one of the States wherein he shall have resided for the term of one year at least, and making proof to the satisfaction of such Court that he is a person of good character, and taking the oath or affirmation prescribed by law to support the Constitution of the United States, which Oath or Affirmation such Court shall administer, and the Clerk of such Court shall record such Application, and the proceedings thereon; and thereupon such person shall be considered as a Citizen of the United States.
It's a toss-up - NB Pro (and NB2) were in the lead for a long time, but gpt-image-1.5 and then particularly gpt-image-2 pretty much closed the gap on my GenAI Image Showdown benchmark site.
NB still generates better looking images though for the most part - gpt-image series is still affected by the yellow saturation issues though its been heavily mitigated.
gpt-image-2 still has a very visible banding/artifacting, but nb2 and gpti2 each have their strengths to the point that its momentarily worth running the same prompt through both and grabbing the better result, then feeding it to the opposite model to tweak it. both of them do a better job of not "getting stuck" regurgitating the same thing over and over when the base image comes from the opposite model. i tend to try and limit how many edits i make with gpti2 because successive iterations degrade the image much much faster.
Apple may be behind, and even getting sued for false advertising around AI features, but at least they haven’t spent hundreds of billions of dollars with no indication of how they’ll make their money back.
You’re right, they’re simply playing a different game. That said, Apple sold millions of phones with the promise that 3 months later users would be able to use AI to automate their phones and use Siri similar to how they use ChatGPT. That was summer 2024, and it still hasn’t shipped.
Fair, many tools trade off cost for performance/adherence. So far our experience is good with Semble (at least with Anthropic/OpenAI models), but if you have any feedback feel free to reach out. We're working on evaluations for this as well, though it will take some time as this is much harder than benchmarking retrieval quality.
reply