I think that software engineers need to start strongly considering a professional organization/association akin to Drs and Lawyers.
Do you think that other professionals would put up with this shit? Absolutely not.
There’s also the matter of an association binding its members by a professional code of ethics. The software engineers who are tasked with building systems like these need to stand up and say no when asked to do. These systems will make life worse for you, your coworkers, and society.
> Do you think that other professionals would put up with this shit? Absolutely not.
Lots of them do?
>> very employee at JPMC has a profile in the WADU database.
This, allegedly, isnt just for swes. JPMC has legal staff who would be, "putting up with this shit" as well as other countless employees in other orgs who would have the same type of monitoring.
If anything engineers are technical enough to at least know their work computer is being monitored. Many employees dont even know that IT can likely access their computer at any time.
I guess I should have clarified, not put up with being monitored but, more to my second paragraph, be so willing and eager (maybe even write a fun technical blog about it) to construct cages for themselves and others.
My understanding of LLMs is sub-par at best, could someone explain where the randomness comes from in the event that the model temperature is 0?
I guess I was imagining that if temperature was 0, and the model was not being continuously trained, the weights wouldn’t change, and the output would be deterministic.
Is this a feature of LLMs more generally or has OpenAI more specifically introduced some other degree of randomness in their models?
It's not the LLM, but the hardware. GPU operations generally involve concurrency that makes them non-deterministic, unless you give up some speed to make them deterministic.
Specifically, as I ubderstand it, the accumulation of rounding errors differs with the order in which floating point values are completed and intermediate aggregates are calculated, unless you put wait conditions in so that the aggregation order is fixed even if the completion order varies, which reduces efficient use of available compute cores in exchange for determinism.
Interestingly a few states do offer a Codeine cough syrup without a prescription[0].
From the article: “Codeine regulation varies widely across countries and states. In some states, including North Carolina, Florida, Oklahoma, and Iowa,1,2 pharmacists can dispense a small quantity of cough syrup containing codeine OTC. Based on individual state law, this information may not be logged into the prescription drug monitoring program, which may contribute to providers being unaware of their patients’ use of codeine.”
This is an exceedingly rare practice in the US however.
I think that your first few points are fair, but I was a little confused about #4.
I had never heard of this, and it seemed important, some cursory research turned up many twitter posts from individuals amplifying your version of events.
I also found a write up on the situation from TechDirt[0]. The article is fairly good and well sourced, but it paints a substantially different picture than what you describe.
"31. After the 2020 election, when EIP was renamed the Virality Project, the Stanford lab was on-boarded to Twitter’s JIRA ticketing system, absorbing this government proxy into Twitter infrastructure – with a capability of taking in an incredible 50 million tweets a day."
"It’s crucial to reiterate: EIP was partnered with state entities like CISA and GEC while seeking elimination of millions of tweets. In the #Twitter Files, Twitter execs did not distinguish between organizations, using phrases like ‘According to CIS[A], escalated via EIP,’" Taibbi wrote. "After the 2020 election, when EIP was renamed the Virality Project, the Stanford lab was on-boarded to Twitter’s JIRA ticketing system, absorbing this government proxy into Twitter infrastructure – with a capability of taking in an incredible 50 million tweets a day.
Definition: Many arguments rely on an analogy between two or more objects, ideas, or situations. If the two things that are being compared aren’t really alike in the relevant respects, the analogy is a weak one, and the argument that relies on it commits the fallacy of weak analogy.
Tip: Identify what properties are important to the claim you’re making, and see whether the two things you’re comparing both share those properties.“[1]
I suspect the poster you are responding to disagrees that it is a relevant comparison, since commodities, their value, and subsequently the way in which they are traded is in some way influenced by their “vital” civilizational importance.
The implication here is that crypto does not share that crucial (for the analogy) property of vital importance, and as such the trading patterns/market forces are not comparable.
In this case is it reasonable to point out that the analogy is weak, we do not have to ignore how the two candidate comparables are different if they differ in fundamental ways.
Do you think that other professionals would put up with this shit? Absolutely not.
There’s also the matter of an association binding its members by a professional code of ethics. The software engineers who are tasked with building systems like these need to stand up and say no when asked to do. These systems will make life worse for you, your coworkers, and society.
Stop trading your conscience for cash.