richardjennings

5 ms·

1.
by richardjennings 1mo ago
If you do not provide access to tools the LLM cannot do anything other than generate tokens. So really it is not about sandboxing a LLM but more about having control over what tools can be accessed and what they can do. Tools can be sandbox
2.
by richardjennings 3mo ago
This goes one of two ways. Either the LLM public capability is not sufficient to positively contribute, or it is. If it is not sufficient to positively contribute, open source projects become drowned in low quality contributions. If it is s
3.
by richardjennings 6mo ago
/loop message ping every 4 minutes keeps the cache warm while the CC REPL is not active.
4.
by richardjennings 6mo ago
> The shorter "thinking" is, the less is the probability of it going astray As long as the error introduced by more steps is less than the compounding error of sub-optimal token sampling, I would expect a better result. I think
5.
by richardjennings 6mo ago
I think over-thinking is only solved by thinking more, not less. This is only viable once some intelligence threshold is reached, which I think Anthropic has borderline achieved.
6.
by richardjennings 6mo ago
You cannot control the effort setting sub-agents use and you also cannot use /effort max as a default (outside of using an alias).
7.
by richardjennings 6mo ago
I was not aware the default effort had changed to medium until the quality of output nosedived. This cost me perhaps a day of work to rectify. I now ensure effort is set to max and have not had a terrible session since. Please may I have a
8.
by richardjennings 6mo ago
You can take DiskANN using OPQ and Vamana and get near zero indexing time and better recall using TurboQuant ... Nice !
9.
by richardjennings 8mo ago
This is similar to what I do. I instruct an Architect mode with a set of rules related to phased implementation and detailed code artifacts output to a report.md file. After a couple of rounds of review and usually some responses that eithe
10.
by richardjennings 9mo ago
SOTA LLMs are now quite good at typing out code that passes tests. If you are able to instruct the creation of sufficient tests and understand the code generated structurally, there is a significant multiplier in productivity. I have found
11.
by richardjennings 10mo ago
Indeed. The suggestion is that classical reality cannot exist without absolute values (an instantiation if you will), consciousness operates where those values have not been established, therefore consciousness is free from classical determ
12.
Thought Experiment: Measuring the Circumference of the Earth from a Bedroom
3 points by richardjennings 10mo ago|2 comments
13.
by richardjennings 1y ago
Is it possible that LLMs could challenge Data Compression Information theory ? Reading this made me wonder how much can be inferred via understanding and thus removed from the minimal necessary representation.
14.
by richardjennings 1y ago
I very much agree with the Author. "Do not reinvent the wheel" is unfortunate in that it muddles several things into one piece of subjective advice. The advice as I would give it is: "Try to re-invent the things you are inter
15.
by richardjennings 2y ago
> The £12.50 charge is for old, polluting cars. It can apply to cars with a £0 VED that were built less than 10 years ago.
16.
by richardjennings 2y ago
It is possible to include CPython in a CGO program - allowing Python to be executed from within the CGO process directly. This comes with some complexities - GIL and thread safety in Go routines, complexity of cross-compiling between archit
17.
by richardjennings 2y ago
I had a reasonable experience with Diagrams previously. I have always been somewhat of a sucker for the theoretical potential for diagrams as code. As a Diagram is generated programatically - it can also be generated dynamically - for examp
18.
Ask HN: Are Open Source Projects Products?
5 points by richardjennings 2y ago|11 comments
19.
by richardjennings 2y ago
I scrape power usage metrics from Tapo P110s and push them to Grafana Cloud using https://github.com/richardjennings/tapmon - although as other commenters have noted - using Wifi for smart plugs has its rough edges.
20.
by richardjennings 3y ago
I have learnt to withhold my natural tendency to care too much, or invest too much of myself personally into my work. When I invest too much personally, I am too exposed emotionally and I run the risk of burning out.
21.
by richardjennings 3y ago
The line of reasoning seems valid, do what you do best; anyone who has had to deal with stored procedures in legacy systems will however disagree. Store procedures are to be avoided.
22.
by richardjennings 4y ago
Agreed. Knowing the opportunity cost must weigh heavily. Bytes and Food. The new boots and cats.
23.
by richardjennings 4y ago
Can I ask if you might consider leveraging your position to aid a charity ?
24.
by richardjennings 4y ago
TOTP is a popular MFA mechanism that is composed from a shared secret and settings (often communicated by qr code). Utilities like LastPass can generate TOTP codes for you if you share the config. The statement reads like any MFA config sha
25.
by richardjennings 4y ago
I think you are right. Kubernetes can be a fairly exotic solution to fairly mundane problems. There is however, perhaps, an inherent elegance in API design and separation of concerns which could result in a transformation in the control of
26.
by richardjennings 4y ago
Broken gets fixed; Shoddy lasts forever.
27.
Show HN: GitOps Repository Management(github.com)
2 points by richardjennings 4y ago|0 comments
28.
by richardjennings 5y ago
I have been using Lima https://github.com/lima-vm/lima which incidentally is the project behind both rancher desktop and colima I believe. QEMU support allows for running e.g. x86 on arm (with perf/efficiency trad
29.
by richardjennings 5y ago
Not suitable for children. I sent my daughter the link as I thought she might find it interesting. The word she got was "wankerout" which was read out loud, unexpected and not appropriate.
30.
by richardjennings 5y ago
I believe that the BBC can provide value for money and remain relevant if and only if there is an internal recognition that any competition they face as an organisation is against the best social value that a public broadcasting utility can

More ›