@malikules There are other LLMs. Here is a prompt on the topic & the beginning of response. There were many screens of info I left out, but here's the summary.
@codypd In-motion Starlink plans for ground/marine use have been limited to 100mph since March to keep people from using low price plans in planes when aviation use began expanding. https://t.co/Ty9z0IocMk
@HayekRonald They are also likely purchasing non-aviation licenses. In-motion plans for ground use were limited back in March to separate the growing aviation use that have expensive plans. 100mph limit was implemented.https://t.co/Ty9z0IocMk
@ATinyGreenCell Yes. Industry often blocks publication. Also employment contracts w things like anything you create even thru collaboration for 1yr after leaving are assumed to have been created during employment and are property of employer. It gums up options for a time.
@iskander My career has been mostly vaccines so the most useful applications fo new tools become annoying and my oncology interests include oncoviral therapies. Investigating approaches to avoid replication in non-tumor cells is bad?
@alexcovo_eth@bot I use the Grok chat for items outside scope of my bots and to create complex prompts I want to give to a bot. Delete one offs that you dont need in order to cleanup context space. I rarely code outside CLI but I use chat to design PRD and Build prompt.
@simonmaechling I agree. Binding isn't neutralization. Trad route is ID binding mAbs from ELISAs of indiv wells of stacks of plates then look for neutralizing clones among the binders. Current method for anti-venom is inject horses, cleanup serum, bottle.
@parmita Im not a fan of nondeterministic LLMs using "skills" that are just the ML models you used to use . Best workflows will be same as your PreciBot. A deterministic algo that doesn't risk chg. Design of the algos will be done by leveraging LLMs. QA is important
@FischerKing64 US also has to compete w alternatives like a higher return loaning the money to Google. I dont see 30 basis pts as being a big difference in that light.
@LordoftheDrugs@Prof_Oak_ I believe the point is if one is saturated, the rest is free to bind the other target, leading to a higher degree of saturation in the unsaturated target. This is a case where you don't require obligate dual binding.
@dshahh A box of meal recipes in my kitchen could be very relevant to my health, but frequency of use of ea is important and not discernible from making a copy of my recipe box.
@ATinyGreenCell I also have a lg orchid collection so I have a bot for that longevity. One task was monitoring the weather for optimal time to bring which orchids inside w daily updates on date chgs due to forecast. Dates chgd for Vandas but not for phragmipediums.
@ATinyGreenCell For coding, I typically use CLI. Agents beat chat when you want MoE advantage, or to separate, narrow and retain project specific goals. Also for ongoing tasks. My computational immunology agent is very targeted and knows my projects.